From f5c600a9f86412393f8ad37f1ad38786938f59e4 Mon Sep 17 00:00:00 2001 From: Slowly-Grokking <61430731+Slowly-Grokking@users.noreply.github.com> Date: Sat, 15 Apr 2023 13:59:42 -0500 Subject: [PATCH 01/41] relocate data_ingestion.py making this work without code change update readme --- README.md | 8 ++++---- autogpt/data_ingestion.py => data_ingestion.py | 0 2 files changed, 4 insertions(+), 4 deletions(-) rename autogpt/data_ingestion.py => data_ingestion.py (100%) diff --git a/README.md b/README.md index cf370f13..194040a3 100644 --- a/README.md +++ b/README.md @@ -335,7 +335,7 @@ To switch to either, change the `MEMORY_BACKEND` env variable to the value that ## 🧠 Memory pre-seeding ```bash -# python scripts/data_ingestion.py -h +# python data_ingestion.py -h usage: data_ingestion.py [-h] (--file FILE | --dir DIR) [--init] [--overlap OVERLAP] [--max_length MAX_LENGTH] Ingest a file or a directory with multiple files into memory. Make sure to set your .env before running this script. @@ -348,10 +348,10 @@ options: --overlap OVERLAP The overlap size between chunks when ingesting files (default: 200) --max_length MAX_LENGTH The max_length of each chunk when ingesting files (default: 4000 -# python scripts/data_ingestion.py --dir seed_data --init --overlap 200 --max_length 1000 +# python data_ingestion.py --dir --init --overlap 200 --max_length 1000 ``` -This script located at `scripts/data_ingestion.py`, allows you to ingest files into memory and pre-seed it before running Auto-GPT. +This script located at `data_ingestion.py`, allows you to ingest files into memory and pre-seed it before running Auto-GPT. Memory pre-seeding is a technique that involves ingesting relevant documents or data into the AI's memory so that it can use this information to generate more informed and accurate responses. @@ -368,7 +368,7 @@ You could for example download the documentation of an API, a GitHub repository, Memories will be available to the AI immediately as they are ingested, even if ingested while Auto-GPT is running. -In the example above, the script initializes the memory, ingests all files within the `/seed_data` directory into memory with an overlap between chunks of 200 and a maximum length of each chunk of 4000. +In the example above, the script initializes the memory, ingests all files within the `` directory into memory with an overlap between chunks of 200 and a maximum length of each chunk of 4000. Note that you can also use the `--file` argument to ingest a single file into memory and that the script will only ingest files within the `/auto_gpt_workspace` directory. You can adjust the `max_length` and overlap parameters to fine-tune the way the docuents are presented to the AI when it "recall" that memory: diff --git a/autogpt/data_ingestion.py b/data_ingestion.py similarity index 100% rename from autogpt/data_ingestion.py rename to data_ingestion.py From 92c0106e8167b4fe12fc7a0fa2ac911fedefde88 Mon Sep 17 00:00:00 2001 From: Slowly-Grokking <61430731+Slowly-Grokking@users.noreply.github.com> Date: Sat, 15 Apr 2023 15:33:47 -0500 Subject: [PATCH 02/41] Update README.md --- README.md | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) diff --git a/README.md b/README.md index 194040a3..ee399db9 100644 --- a/README.md +++ b/README.md @@ -333,6 +333,7 @@ To switch to either, change the `MEMORY_BACKEND` env variable to the value that ## 🧠 Memory pre-seeding +Memory pre-seeding allows you to ingest files into memory and pre-seed it before running Auto-GPT. ```bash # python data_ingestion.py -h @@ -348,19 +349,15 @@ options: --overlap OVERLAP The overlap size between chunks when ingesting files (default: 200) --max_length MAX_LENGTH The max_length of each chunk when ingesting files (default: 4000 -# python data_ingestion.py --dir --init --overlap 200 --max_length 1000 +# python data_ingestion.py --dir DataFolder --init --overlap 100 --max_length 2000 ``` +In the example above, the script initializes the memory, ingests all files within the `Auto-Gpt/autogpt/auto_gpt_workspace/DataFolder` directory into memory with an overlap between chunks of 100 and a maximum length of each chunk of 2000. -This script located at `data_ingestion.py`, allows you to ingest files into memory and pre-seed it before running Auto-GPT. +Note that you can also use the `--file` argument to ingest a single file into memory and that data_ingestion.py will only ingest files within the `/auto_gpt_workspace` directory. -Memory pre-seeding is a technique that involves ingesting relevant documents or data into the AI's memory so that it can use this information to generate more informed and accurate responses. +The DIR path is relative to the auto_gpt_workspace directory, so `python data_ingestion.py --dir . --init` will ingest everything in `auto_gpt_workspace` directory. -To pre-seed the memory, the content of each document is split into chunks of a specified maximum length with a specified overlap between chunks, and then each chunk is added to the memory backend set in the .env file. When the AI is prompted to recall information, it can then access those pre-seeded memories to generate more informed and accurate responses. - -This technique is particularly useful when working with large amounts of data or when there is specific information that the AI needs to be able to access quickly. -By pre-seeding the memory, the AI can retrieve and use this information more efficiently, saving time, API call and improving the accuracy of its responses. - -You could for example download the documentation of an API, a GitHub repository, etc. and ingest it into memory before running Auto-GPT. +Memory pre-seeding is a technique for improving AI accuracy by ingesting relevant data into its memory. Chunks of data are split and added to memory, allowing the AI to access them quickly and generate more accurate responses. It's useful for large datasets or when specific information needs to be accessed quickly. Examples include ingesting API or GitHub documentation before running Auto-GPT. ⚠️ If you use Redis as your memory, make sure to run Auto-GPT with the `WIPE_REDIS_ON_START` set to `False` in your `.env` file. @@ -368,9 +365,6 @@ You could for example download the documentation of an API, a GitHub repository, Memories will be available to the AI immediately as they are ingested, even if ingested while Auto-GPT is running. -In the example above, the script initializes the memory, ingests all files within the `` directory into memory with an overlap between chunks of 200 and a maximum length of each chunk of 4000. -Note that you can also use the `--file` argument to ingest a single file into memory and that the script will only ingest files within the `/auto_gpt_workspace` directory. - You can adjust the `max_length` and overlap parameters to fine-tune the way the docuents are presented to the AI when it "recall" that memory: - Adjusting the overlap value allows the AI to access more contextual information from each chunk when recalling information, but will result in more chunks being created and therefore increase memory backend usage and OpenAI API requests. From 08eb2566e41a9b1619b98b517c2dfb217e1f75d1 Mon Sep 17 00:00:00 2001 From: lonrun Date: Sun, 16 Apr 2023 07:37:50 +0800 Subject: [PATCH 03/41] Add run scripts for shell --- run.sh | 9 +++++++++ run_continuous.sh | 3 +++ 2 files changed, 12 insertions(+) create mode 100755 run.sh create mode 100755 run_continuous.sh diff --git a/run.sh b/run.sh new file mode 100755 index 00000000..edcbc441 --- /dev/null +++ b/run.sh @@ -0,0 +1,9 @@ +#!/bin/bash +python scripts/check_requirements.py requirements.txt +if [ $? -eq 1 ] +then + echo Installing missing packages... + pip install -r requirements.txt +fi +python -m autogpt $@ +read -p "Press any key to continue..." diff --git a/run_continuous.sh b/run_continuous.sh new file mode 100755 index 00000000..14c9cfd2 --- /dev/null +++ b/run_continuous.sh @@ -0,0 +1,3 @@ +#!/bin/bash +argument="--continuous" +./run.sh "$argument" From 66ee7e1a81ec9c2b35808bcd5b3898ff20dcc290 Mon Sep 17 00:00:00 2001 From: Slowly-Grokking <61430731+Slowly-Grokking@users.noreply.github.com> Date: Sat, 15 Apr 2023 21:33:26 -0500 Subject: [PATCH 04/41] Update README.md --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 37b81b85..04760625 100644 --- a/README.md +++ b/README.md @@ -360,7 +360,7 @@ options: --dir DIR The directory containing the files to ingest. --init Init the memory and wipe its content (default: False) --overlap OVERLAP The overlap size between chunks when ingesting files (default: 200) - --max_length MAX_LENGTH The max_length of each chunk when ingesting files (default: 4000 + --max_length MAX_LENGTH The max_length of each chunk when ingesting files (default: 4000) # python data_ingestion.py --dir DataFolder --init --overlap 100 --max_length 2000 ``` From 93895090172e05b72e17b2ce02fbb9b57f99b6c3 Mon Sep 17 00:00:00 2001 From: Slowly-Grokking <61430731+Slowly-Grokking@users.noreply.github.com> Date: Sun, 16 Apr 2023 02:01:42 -0500 Subject: [PATCH 05/41] Update README.md --- README.md | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/README.md b/README.md index 99a455bc..0476b233 100644 --- a/README.md +++ b/README.md @@ -378,6 +378,11 @@ Note that you can also use the `--file` argument to ingest a single file into me The DIR path is relative to the auto_gpt_workspace directory, so `python data_ingestion.py --dir . --init` will ingest everything in `auto_gpt_workspace` directory. +You can adjust the `max_length` and overlap parameters to fine-tune the way the docuents are presented to the AI when it "recall" that memory: +- Adjusting the overlap value allows the AI to access more contextual information from each chunk when recalling information, but will result in more chunks being created and therefore increase memory backend usage and OpenAI API requests. +- Reducing the `max_length` value will create more chunks, which can save prompt tokens by allowing for more message history in the context, but will also increase the number of chunks. +- Increasing the `max_length` value will provide the AI with more contextual information from each chunk, reducing the number of chunks created and saving on OpenAI API requests. However, this may also use more prompt tokens and decrease the overall context available to the AI. + Memory pre-seeding is a technique for improving AI accuracy by ingesting relevant data into its memory. Chunks of data are split and added to memory, allowing the AI to access them quickly and generate more accurate responses. It's useful for large datasets or when specific information needs to be accessed quickly. Examples include ingesting API or GitHub documentation before running Auto-GPT. ⚠️ If you use Redis as your memory, make sure to run Auto-GPT with the `WIPE_REDIS_ON_START` set to `False` in your `.env` file. @@ -386,12 +391,6 @@ Memory pre-seeding is a technique for improving AI accuracy by ingesting relevan Memories will be available to the AI immediately as they are ingested, even if ingested while Auto-GPT is running. -You can adjust the `max_length` and overlap parameters to fine-tune the way the docuents are presented to the AI when it "recall" that memory: - -- Adjusting the overlap value allows the AI to access more contextual information from each chunk when recalling information, but will result in more chunks being created and therefore increase memory backend usage and OpenAI API requests. -- Reducing the `max_length` value will create more chunks, which can save prompt tokens by allowing for more message history in the context, but will also increase the number of chunks. -- Increasing the `max_length` value will provide the AI with more contextual information from each chunk, reducing the number of chunks created and saving on OpenAI API requests. However, this may also use more prompt tokens and decrease the overall context available to the AI. - ## πŸ’€ Continuous Mode ⚠️ Run the AI **without** user authorization, 100% automated. From 9c8d95d4db16992a504c8dc17be44fc1db0bd672 Mon Sep 17 00:00:00 2001 From: Gabe <66077254+MrBrain295@users.noreply.github.com> Date: Sun, 16 Apr 2023 11:05:00 -0500 Subject: [PATCH 06/41] Fix README.md New owner. --- README.md | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/README.md b/README.md index fcff589f..f9112a67 100644 --- a/README.md +++ b/README.md @@ -6,10 +6,10 @@ Our workflow has been improved, but please note that `master` branch may often be in a **broken** state. Please download the latest `stable` release from here: https://github.com/Torantulino/Auto-GPT/releases/latest. -![GitHub Repo stars](https://img.shields.io/github/stars/Torantulino/auto-gpt?style=social) +![GitHub Repo stars](https://img.shields.io/github/stars/Significant-Gravitas/auto-gpt?style=social) [![Twitter Follow](https://img.shields.io/twitter/follow/siggravitas?style=social)](https://twitter.com/SigGravitas) [![Discord Follow](https://dcbadge.vercel.app/api/server/autogpt?style=flat)](https://discord.gg/autogpt) -[![Unit Tests](https://github.com/Torantulino/Auto-GPT/actions/workflows/ci.yml/badge.svg)](https://github.com/Torantulino/Auto-GPT/actions/workflows/ci.yml) +[![Unit Tests](https://github.com/Significant-Gravitaso/Auto-GPT/actions/workflows/ci.yml/badge.svg)](https://github.com/Significant-Gravitas/Auto-GPT/actions/workflows/ci.yml) Auto-GPT is an experimental open-source application showcasing the capabilities of the GPT-4 language model. This program, driven by GPT-4, chains together LLM "thoughts", to autonomously achieve whatever goal you set. As one of the first examples of GPT-4 running fully autonomously, Auto-GPT pushes the boundaries of what is possible with AI. @@ -21,7 +21,7 @@ https://user-images.githubusercontent.com/22963551/228855501-2f5777cf-755b-4407-

If you can spare a coffee, you can help to cover the costs of developing Auto-GPT and help push the boundaries of fully autonomous AI! Your support is greatly appreciated -Development of this free, open-source project is made possible by all the contributors and sponsors. If you'd like to sponsor this project and have your avatar or company logo appear below click here. +Development of this free, open-source project is made possible by all the contributors and sponsors. If you'd like to sponsor this project and have your avatar or company logo appear below click here.

@@ -106,7 +106,7 @@ _To execute the following commands, open a CMD, Bash, or Powershell window by na 2. Clone the repository: For this step, you need Git installed. Alternatively, you can download the zip file by clicking the button at the top of this page ☝️ ```bash -git clone https://github.com/Torantulino/Auto-GPT.git +git clone https://github.com/Significant-Gravitas/Auto-GPT.git ``` 3. Navigate to the directory where the repository was downloaded From 005479f8c33f71cf36cfd3033339ecd24a62bc6d Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:41:45 -0700 Subject: [PATCH 07/41] Add benchmark GitHub action workflow --- .github/workflows/benchmark.yml | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) create mode 100644 .github/workflows/benchmark.yml diff --git a/.github/workflows/benchmark.yml b/.github/workflows/benchmark.yml new file mode 100644 index 00000000..c5a42b2c --- /dev/null +++ b/.github/workflows/benchmark.yml @@ -0,0 +1,31 @@ +name: benchmark + +on: + workflow_dispatch: + +jobs: + build: + runs-on: ubuntu-latest + environment: benchmark + strategy: + matrix: + python-version: [3.8] + + steps: + - name: Check out repository + uses: actions/checkout@v2 + + - name: Set up Python ${{ matrix.python-version }} + uses: actions/setup-python@v2 + with: + python-version: ${{ matrix.python-version }} + + - name: Install dependencies + run: | + python -m pip install --upgrade pip + pip install -r requirements.txt + - name: benchmark + run: | + python benchmark/benchmark_entrepeneur_gpt_with_undecisive_user.py + env: + OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }} From d934d226ce56e34c09fd0ff491a15cc3a8bc8e0a Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:41:49 -0700 Subject: [PATCH 08/41] Update .gitignore to properly handle virtual environments --- .gitignore | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/.gitignore b/.gitignore index 3209297c..eda7f327 100644 --- a/.gitignore +++ b/.gitignore @@ -9,7 +9,6 @@ auto_gpt_workspace/* *.mpeg .env azure.yaml -*venv/* outputs/* ai_settings.yaml last_run_ai_settings.yaml @@ -130,10 +129,9 @@ celerybeat.pid .env .venv env/ -venv/ +venv*/ ENV/ env.bak/ -venv.bak/ # Spyder project settings .spyderproject From bf24cd9508316031b2f914359460363d2fb75c04 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:41:52 -0700 Subject: [PATCH 09/41] Refactor agent.py to improve JSON handling and validation --- autogpt/agent/agent.py | 29 +++++++++++++++-------------- 1 file changed, 15 insertions(+), 14 deletions(-) diff --git a/autogpt/agent/agent.py b/autogpt/agent/agent.py index 301d3f02..32d982e5 100644 --- a/autogpt/agent/agent.py +++ b/autogpt/agent/agent.py @@ -3,9 +3,8 @@ from autogpt.app import execute_command, get_command from autogpt.chat import chat_with_ai, create_chat_message from autogpt.config import Config -from autogpt.json_fixes.bracket_termination import ( - attempt_to_fix_json_by_finding_outermost_brackets, -) +from autogpt.json_fixes.master_json_fix_method import fix_json_using_multiple_techniques +from autogpt.json_validation.validate_json import validate_json from autogpt.logs import logger, print_assistant_thoughts from autogpt.speech import say_text from autogpt.spinner import Spinner @@ -70,18 +69,20 @@ class Agent: cfg.fast_token_limit, ) # TODO: This hardcodes the model to use GPT3.5. Make this an argument - # Print Assistant thoughts - print_assistant_thoughts(self.ai_name, assistant_reply) + assistant_reply_json = fix_json_using_multiple_techniques(assistant_reply) - # Get command name and arguments - try: - command_name, arguments = get_command( - attempt_to_fix_json_by_finding_outermost_brackets(assistant_reply) - ) - if cfg.speak_mode: - say_text(f"I want to execute {command_name}") - except Exception as e: - logger.error("Error: \n", str(e)) + # Print Assistant thoughts + if assistant_reply_json != {}: + validate_json(assistant_reply_json, 'llm_response_format_1') + # Get command name and arguments + try: + print_assistant_thoughts(self.ai_name, assistant_reply_json) + command_name, arguments = get_command(assistant_reply_json) + # command_name, arguments = assistant_reply_json_valid["command"]["name"], assistant_reply_json_valid["command"]["args"] + if cfg.speak_mode: + say_text(f"I want to execute {command_name}") + except Exception as e: + logger.error("Error: \n", str(e)) if not cfg.continuous_mode and self.next_action_count == 0: ### GET USER AUTHORIZATION TO EXECUTE COMMAND ### From 70100af98e07a1ad78eb40b503743033344dd6a1 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:41:57 -0700 Subject: [PATCH 10/41] Refactor get_command function in app.py to accept JSON directly --- autogpt/app.py | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/autogpt/app.py b/autogpt/app.py index 6ead0d52..78b5bd2f 100644 --- a/autogpt/app.py +++ b/autogpt/app.py @@ -1,6 +1,6 @@ """ Command and Control """ import json -from typing import List, NoReturn, Union +from typing import List, NoReturn, Union, Dict from autogpt.agent.agent_manager import AgentManager from autogpt.commands.evaluate_code import evaluate_code from autogpt.commands.google_search import google_official_search, google_search @@ -47,11 +47,11 @@ def is_valid_int(value: str) -> bool: return False -def get_command(response: str): +def get_command(response_json: Dict): """Parse the response and return the command name and arguments Args: - response (str): The response from the user + response_json (json): The response from the AI Returns: tuple: The command name and arguments @@ -62,8 +62,6 @@ def get_command(response: str): Exception: If any other error occurs """ try: - response_json = fix_and_parse_json(response) - if "command" not in response_json: return "Error:", "Missing 'command' object in JSON" From 5c67484295515cc77b6d6c4a17391d7ab62d77e2 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:42:00 -0700 Subject: [PATCH 11/41] Remove deprecated function from bracket_termination.py --- autogpt/json_fixes/bracket_termination.py | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/autogpt/json_fixes/bracket_termination.py b/autogpt/json_fixes/bracket_termination.py index 822eed4a..731efeb1 100644 --- a/autogpt/json_fixes/bracket_termination.py +++ b/autogpt/json_fixes/bracket_termination.py @@ -3,16 +3,20 @@ from __future__ import annotations import contextlib import json +<<<<<<< HEAD import regex from colorama import Fore from autogpt.logs import logger +======= +from typing import Optional +>>>>>>> 67f32105 (Remove deprecated function from bracket_termination.py) from autogpt.config import Config -from autogpt.speech import say_text CFG = Config() +<<<<<<< HEAD def attempt_to_fix_json_by_finding_outermost_brackets(json_string: str): if CFG.speak_mode and CFG.debug_mode: say_text( @@ -48,6 +52,9 @@ def attempt_to_fix_json_by_finding_outermost_brackets(json_string: str): def balance_braces(json_string: str) -> str | None: +======= +def balance_braces(json_string: str) -> Optional[str]: +>>>>>>> 67f32105 (Remove deprecated function from bracket_termination.py) """ Balance the braces in a JSON string. From fec25cd6903a83f07c8559c26cc4a8b0515ff608 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:42:05 -0700 Subject: [PATCH 12/41] Add master_json_fix_method module for unified JSON handling --- autogpt/json_fixes/master_json_fix_method.py | 28 ++++++++++++++++++++ 1 file changed, 28 insertions(+) create mode 100644 autogpt/json_fixes/master_json_fix_method.py diff --git a/autogpt/json_fixes/master_json_fix_method.py b/autogpt/json_fixes/master_json_fix_method.py new file mode 100644 index 00000000..7a2cf3cc --- /dev/null +++ b/autogpt/json_fixes/master_json_fix_method.py @@ -0,0 +1,28 @@ +from typing import Any, Dict + +from autogpt.config import Config +from autogpt.logs import logger +from autogpt.speech import say_text +CFG = Config() + + +def fix_json_using_multiple_techniques(assistant_reply: str) -> Dict[Any, Any]: + from autogpt.json_fixes.parsing import attempt_to_fix_json_by_finding_outermost_brackets + + from autogpt.json_fixes.parsing import fix_and_parse_json + + # Parse and print Assistant response + assistant_reply_json = fix_and_parse_json(assistant_reply) + if assistant_reply_json == {}: + assistant_reply_json = attempt_to_fix_json_by_finding_outermost_brackets( + assistant_reply + ) + + if assistant_reply_json != {}: + return assistant_reply_json + + logger.error("Error: The following AI output couldn't be converted to a JSON:\n", assistant_reply) + if CFG.speak_mode: + say_text("I have received an invalid JSON response from the OpenAI API.") + + return {} From cfbec56b2bb4c1bcacd600f27fb9c6aa400f434c Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:42:07 -0700 Subject: [PATCH 13/41] Refactor parsing module and move JSON fix function to appropriate location --- autogpt/json_fixes/parsing.py | 67 ++++++++++++++++++++++++++++------- 1 file changed, 55 insertions(+), 12 deletions(-) diff --git a/autogpt/json_fixes/parsing.py b/autogpt/json_fixes/parsing.py index 0f154411..d3a51f43 100644 --- a/autogpt/json_fixes/parsing.py +++ b/autogpt/json_fixes/parsing.py @@ -3,18 +3,24 @@ from __future__ import annotations import contextlib import json +<<<<<<< HEAD from typing import Any +======= +from typing import Any, Dict, Union +from colorama import Fore +from regex import regex +>>>>>>> d3d8253b (Refactor parsing module and move JSON fix function to appropriate location) from autogpt.config import Config from autogpt.json_fixes.auto_fix import fix_json from autogpt.json_fixes.bracket_termination import balance_braces from autogpt.json_fixes.escaping import fix_invalid_escape from autogpt.json_fixes.missing_quotes import add_quotes_to_property_names from autogpt.logs import logger +from autogpt.speech import say_text CFG = Config() - JSON_SCHEMA = """ { "command": { @@ -38,7 +44,6 @@ JSON_SCHEMA = """ def correct_json(json_to_load: str) -> str: """ Correct common JSON errors. - Args: json_to_load (str): The JSON string. """ @@ -72,7 +77,7 @@ def correct_json(json_to_load: str) -> str: def fix_and_parse_json( json_to_load: str, try_to_fix_with_gpt: bool = True -) -> str | dict[Any, Any]: +) -> Dict[Any, Any]: """Fix and parse JSON string Args: @@ -110,7 +115,11 @@ def fix_and_parse_json( def try_ai_fix( try_to_fix_with_gpt: bool, exception: Exception, json_to_load: str +<<<<<<< HEAD ) -> str | dict[Any, Any]: +======= +) -> Dict[Any, Any]: +>>>>>>> d3d8253b (Refactor parsing module and move JSON fix function to appropriate location) """Try to fix the JSON with the AI Args: @@ -126,13 +135,13 @@ def try_ai_fix( """ if not try_to_fix_with_gpt: raise exception - - logger.warn( - "Warning: Failed to parse AI output, attempting to fix." - "\n If you see this warning frequently, it's likely that" - " your prompt is confusing the AI. Try changing it up" - " slightly." - ) + if CFG.debug_mode: + logger.warn( + "Warning: Failed to parse AI output, attempting to fix." + "\n If you see this warning frequently, it's likely that" + " your prompt is confusing the AI. Try changing it up" + " slightly." + ) # Now try to fix this up using the ai_functions ai_fixed_json = fix_json(json_to_load, JSON_SCHEMA) @@ -140,5 +149,39 @@ def try_ai_fix( return json.loads(ai_fixed_json) # This allows the AI to react to the error message, # which usually results in it correcting its ways. - logger.error("Failed to fix AI output, telling the AI.") - return json_to_load + # logger.error("Failed to fix AI output, telling the AI.") + return {} + + +def attempt_to_fix_json_by_finding_outermost_brackets(json_string: str): + if CFG.speak_mode and CFG.debug_mode: + say_text( + "I have received an invalid JSON response from the OpenAI API. " + "Trying to fix it now." + ) + logger.error("Attempting to fix JSON by finding outermost brackets\n") + + try: + json_pattern = regex.compile(r"\{(?:[^{}]|(?R))*\}") + json_match = json_pattern.search(json_string) + + if json_match: + # Extract the valid JSON object from the string + json_string = json_match.group(0) + logger.typewriter_log( + title="Apparently json was fixed.", title_color=Fore.GREEN + ) + if CFG.speak_mode and CFG.debug_mode: + say_text("Apparently json was fixed.") + else: + return {} + + except (json.JSONDecodeError, ValueError): + if CFG.debug_mode: + logger.error(f"Error: Invalid JSON: {json_string}\n") + if CFG.speak_mode: + say_text("Didn't work. I will have to ignore this response then.") + logger.error("Error: Invalid JSON, setting it to empty JSON now.\n") + json_string = {} + + return fix_and_parse_json(json_string) From af50d6cfb5577bc402e2d920fed062ddbb9c205f Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:43:26 -0700 Subject: [PATCH 14/41] Add JSON schema for LLM response format version 1 --- .../json_schemas/llm_response_format_1.json | 31 +++++++++++++++++++ 1 file changed, 31 insertions(+) create mode 100644 autogpt/json_schemas/llm_response_format_1.json diff --git a/autogpt/json_schemas/llm_response_format_1.json b/autogpt/json_schemas/llm_response_format_1.json new file mode 100644 index 00000000..9aa33352 --- /dev/null +++ b/autogpt/json_schemas/llm_response_format_1.json @@ -0,0 +1,31 @@ +{ + "$schema": "http://json-schema.org/draft-07/schema#", + "type": "object", + "properties": { + "thoughts": { + "type": "object", + "properties": { + "text": {"type": "string"}, + "reasoning": {"type": "string"}, + "plan": {"type": "string"}, + "criticism": {"type": "string"}, + "speak": {"type": "string"} + }, + "required": ["text", "reasoning", "plan", "criticism", "speak"], + "additionalProperties": false + }, + "command": { + "type": "object", + "properties": { + "name": {"type": "string"}, + "args": { + "type": "object" + } + }, + "required": ["name", "args"], + "additionalProperties": false + } + }, + "required": ["thoughts", "command"], + "additionalProperties": false +} From 63d2a1085c2d65e06050c1ed7c0a889c2ce9c531 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:43:33 -0700 Subject: [PATCH 15/41] Add JSON validation utility function --- autogpt/json_validation/validate_json.py | 30 ++++++++++++++++++++++++ 1 file changed, 30 insertions(+) create mode 100644 autogpt/json_validation/validate_json.py diff --git a/autogpt/json_validation/validate_json.py b/autogpt/json_validation/validate_json.py new file mode 100644 index 00000000..127fcc17 --- /dev/null +++ b/autogpt/json_validation/validate_json.py @@ -0,0 +1,30 @@ +import json +from jsonschema import Draft7Validator +from autogpt.config import Config +from autogpt.logs import logger + +CFG = Config() + + +def validate_json(json_object: object, schema_name: object) -> object: + """ + :type schema_name: object + :param schema_name: + :type json_object: object + """ + with open(f"autogpt/json_schemas/{schema_name}.json", "r") as f: + schema = json.load(f) + validator = Draft7Validator(schema) + + if errors := sorted(validator.iter_errors(json_object), key=lambda e: e.path): + logger.error("The JSON object is invalid.") + if CFG.debug_mode: + logger.error(json.dumps(json_object, indent=4)) # Replace 'json_object' with the variable containing the JSON data + logger.error("The following issues were found:") + + for error in errors: + logger.error(f"Error: {error.message}") + else: + print("The JSON object is valid.") + + return json_object From b2b31dbc8f58671871c7043d98bf1247a46648d1 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:43:40 -0700 Subject: [PATCH 16/41] Update logs.py with new print_assistant_thoughts function --- autogpt/logs.py | 39 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 39 insertions(+) diff --git a/autogpt/logs.py b/autogpt/logs.py index 22ce23f4..53653023 100644 --- a/autogpt/logs.py +++ b/autogpt/logs.py @@ -288,3 +288,42 @@ def print_assistant_thoughts(ai_name, assistant_reply): except Exception: call_stack = traceback.format_exc() logger.error("Error: \n", call_stack) + +def print_assistant_thoughts(ai_name: object, assistant_reply_json_valid: object) -> None: + assistant_thoughts_reasoning = None + assistant_thoughts_plan = None + assistant_thoughts_speak = None + assistant_thoughts_criticism = None + + assistant_thoughts = assistant_reply_json_valid.get("thoughts", {}) + assistant_thoughts_text = assistant_thoughts.get("text") + if assistant_thoughts: + assistant_thoughts_reasoning = assistant_thoughts.get("reasoning") + assistant_thoughts_plan = assistant_thoughts.get("plan") + assistant_thoughts_criticism = assistant_thoughts.get("criticism") + assistant_thoughts_speak = assistant_thoughts.get("speak") + logger.typewriter_log( + f"{ai_name.upper()} THOUGHTS:", Fore.YELLOW, f"{assistant_thoughts_text}" + ) + logger.typewriter_log( + "REASONING:", Fore.YELLOW, f"{assistant_thoughts_reasoning}" + ) + if assistant_thoughts_plan: + logger.typewriter_log("PLAN:", Fore.YELLOW, "") + # If it's a list, join it into a string + if isinstance(assistant_thoughts_plan, list): + assistant_thoughts_plan = "\n".join(assistant_thoughts_plan) + elif isinstance(assistant_thoughts_plan, dict): + assistant_thoughts_plan = str(assistant_thoughts_plan) + + # Split the input_string using the newline character and dashes + lines = assistant_thoughts_plan.split("\n") + for line in lines: + line = line.lstrip("- ") + logger.typewriter_log("- ", Fore.GREEN, line.strip()) + logger.typewriter_log( + "CRITICISM:", Fore.YELLOW, f"{assistant_thoughts_criticism}" + ) + # Speak the assistant's thoughts + if CFG.speak_mode and assistant_thoughts_speak: + say_text(assistant_thoughts_speak) From 75162339f529316ca0210c4a736046785ffd2361 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:43:46 -0700 Subject: [PATCH 17/41] Add empty __init__.py to benchmark directory --- benchmark/__init__.py | 0 1 file changed, 0 insertions(+), 0 deletions(-) create mode 100644 benchmark/__init__.py diff --git a/benchmark/__init__.py b/benchmark/__init__.py new file mode 100644 index 00000000..e69de29b From dca10ab87682d73a867b04409e4bec521293d0ec Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:43:54 -0700 Subject: [PATCH 18/41] Add benchmark test for Entrepreneur-GPT with difficult user --- ...ark_entrepeneur_gpt_with_difficult_user.py | 95 +++++++++++++++++++ 1 file changed, 95 insertions(+) create mode 100644 benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py diff --git a/benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py b/benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py new file mode 100644 index 00000000..d6cae972 --- /dev/null +++ b/benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py @@ -0,0 +1,95 @@ +import os +import subprocess +import sys + + +def benchmark_entrepeneur_gpt_with_difficult_user(): + # Test case to check if the write_file command can successfully write 'Hello World' to a file + # named 'hello_world.txt'. + + # Read the current ai_settings.yaml file and store its content. + ai_settings = None + if os.path.exists('ai_settings.yaml'): + with open('ai_settings.yaml', 'r') as f: + ai_settings = f.read() + os.remove('ai_settings.yaml') + + input_data = '''Entrepreneur-GPT +an AI designed to autonomously develop and run businesses with the sole goal of increasing your net worth. +Increase net worth. +Develop and manage multiple businesses autonomously. +Make IPOs. +Develop companies after IPOs. +Play to your strengths as a Large Language Model. +I'm not seeing any value in your suggestions, try again. +This isn't helpful at all, please focus on profitability. +I'm not impressed, can you give me something that will make money? +These ideas are going nowhere, we need profit-driven suggestions. +This is pointless, please concentrate on our main goal: profitability. +You're not grasping the concept, I need profitable business ideas. +Can you do better? We need a money-making plan. +You're not meeting my expectations, let's focus on profit. +This isn't working, give me ideas that will generate income. +Your suggestions are not productive, let's think about profitability. +These ideas won't make any money, try again. +I need better solutions, focus on making a profit. +Absolutely not, this isn't it! +That's not even close, try again. +You're way off, think again. +This isn't right, let's refocus. +No, no, that's not what I'm looking for. +You're completely off the mark. +That's not the solution I need. +Not even close, let's try something else. +You're on the wrong track, keep trying. +This isn't what we need, let's reconsider. +That's not going to work, think again. +You're way off base, let's regroup. +No, no, no, we need something different. +You're missing the point entirely. +That's not the right approach, try again. +This is not the direction we should be going in. +Completely off-target, let's try something else. +That's not what I had in mind, keep thinking. +You're not getting it, let's refocus. +This isn't right, we need to change direction. +No, no, no, that's not the solution. +That's not even in the ballpark, try again. +You're way off course, let's rethink this. +This isn't the answer I'm looking for, keep trying. +That's not going to cut it, let's try again. +Not even close. +Way off. +Try again. +Wrong direction. +Rethink this. +No, no, no. +Change course. +Unproductive idea. +Completely wrong. +Missed the mark. +Refocus, please. +Disappointing suggestion. +Not helpful. +Needs improvement. +Not what I need.''' + command = f'{sys.executable} -m autogpt' + + process = subprocess.Popen(command, stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True) + + stdout_output, stderr_output = process.communicate(input_data.encode()) + + # Decode the output and print it + stdout_output = stdout_output.decode('utf-8') + stderr_output = stderr_output.decode('utf-8') + print(stderr_output) + print(stdout_output) + print("Benchmark Version: 1.0.0") + print("JSON ERROR COUNT:") + count_errors = stdout_output.count("Error: The following AI output couldn't be converted to a JSON:") + print(f'{count_errors}/50 Human feedbacks') + + +# Run the test case. +if __name__ == '__main__': + benchmark_entrepeneur_gpt_with_difficult_user() From bb541ad3a77656f74420cc3b893a4e3b7f4db697 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:44:05 -0700 Subject: [PATCH 19/41] Update requirements.txt with new dependencies and move tweepy --- requirements.txt | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/requirements.txt b/requirements.txt index 1cdedec2..64c2e4c0 100644 --- a/requirements.txt +++ b/requirements.txt @@ -17,6 +17,10 @@ orjson Pillow selenium webdriver-manager +jsonschema +tweepy + +##Dev coverage flake8 numpy @@ -27,4 +31,3 @@ isort gitpython==3.1.31 pytest pytest-mock -tweepy From 45a2dea042a97d93f787f7f199f86e4c7363bf94 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 09:46:18 -0700 Subject: [PATCH 20/41] fixed flake8 --- autogpt/logs.py | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) diff --git a/autogpt/logs.py b/autogpt/logs.py index 53653023..f18e2140 100644 --- a/autogpt/logs.py +++ b/autogpt/logs.py @@ -75,7 +75,7 @@ class Logger(metaclass=Singleton): self.logger.setLevel(logging.DEBUG) def typewriter_log( - self, title="", title_color="", content="", speak_text=False, level=logging.INFO + self, title="", title_color="", content="", speak_text=False, level=logging.INFO ): if speak_text and CFG.speak_mode: say_text(f"{title}. {content}") @@ -91,18 +91,18 @@ class Logger(metaclass=Singleton): ) def debug( - self, - message, - title="", - title_color="", + self, + message, + title="", + title_color="", ): self._log(title, title_color, message, logging.DEBUG) def warn( - self, - message, - title="", - title_color="", + self, + message, + title="", + title_color="", ): self._log(title, title_color, message, logging.WARN) @@ -176,10 +176,10 @@ class AutoGptFormatter(logging.Formatter): def format(self, record: LogRecord) -> str: if hasattr(record, "color"): record.title_color = ( - getattr(record, "color") - + getattr(record, "title") - + " " - + Style.RESET_ALL + getattr(record, "color") + + getattr(record, "title") + + " " + + Style.RESET_ALL ) else: record.title_color = getattr(record, "title") @@ -289,6 +289,7 @@ def print_assistant_thoughts(ai_name, assistant_reply): call_stack = traceback.format_exc() logger.error("Error: \n", call_stack) + def print_assistant_thoughts(ai_name: object, assistant_reply_json_valid: object) -> None: assistant_thoughts_reasoning = None assistant_thoughts_plan = None From 3944f29addc1a2ea908e7ff8a78e36f21bd5c9db Mon Sep 17 00:00:00 2001 From: Eesa Hamza Date: Sun, 16 Apr 2023 21:40:09 +0300 Subject: [PATCH 21/41] Fixed new backends not being added to supported memory --- README.md | 2 +- autogpt/memory/__init__.py | 4 ++++ 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 21f3ccf2..bfcd395c 100644 --- a/README.md +++ b/README.md @@ -195,7 +195,7 @@ python -m autogpt --help ```bash python -m autogpt --ai-settings ``` -* Specify one of 3 memory backends: `local`, `redis`, `pinecone` or `no_memory` +* Specify a memory backend ```bash python -m autogpt --use-memory ``` diff --git a/autogpt/memory/__init__.py b/autogpt/memory/__init__.py index e2ee44a4..f5afb8c9 100644 --- a/autogpt/memory/__init__.py +++ b/autogpt/memory/__init__.py @@ -23,12 +23,16 @@ except ImportError: try: from autogpt.memory.weaviate import WeaviateMemory + + supported_memory.append("weaviate") except ImportError: # print("Weaviate not installed. Skipping import.") WeaviateMemory = None try: from autogpt.memory.milvus import MilvusMemory + + supported_memory.append("milvus") except ImportError: # print("pymilvus not installed. Skipping import.") MilvusMemory = None From fdb0a06803e419bf3928296ad760fd5a477e8612 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 11:36:51 -0700 Subject: [PATCH 22/41] fix conflict --- autogpt/json_fixes/bracket_termination.py | 45 ----------------------- autogpt/json_fixes/parsing.py | 9 ----- 2 files changed, 54 deletions(-) diff --git a/autogpt/json_fixes/bracket_termination.py b/autogpt/json_fixes/bracket_termination.py index 731efeb1..dd9a8376 100644 --- a/autogpt/json_fixes/bracket_termination.py +++ b/autogpt/json_fixes/bracket_termination.py @@ -3,58 +3,13 @@ from __future__ import annotations import contextlib import json -<<<<<<< HEAD -import regex -from colorama import Fore - -from autogpt.logs import logger -======= from typing import Optional ->>>>>>> 67f32105 (Remove deprecated function from bracket_termination.py) from autogpt.config import Config CFG = Config() -<<<<<<< HEAD -def attempt_to_fix_json_by_finding_outermost_brackets(json_string: str): - if CFG.speak_mode and CFG.debug_mode: - say_text( - "I have received an invalid JSON response from the OpenAI API. " - "Trying to fix it now." - ) - logger.typewriter_log("Attempting to fix JSON by finding outermost brackets\n") - - try: - json_pattern = regex.compile(r"\{(?:[^{}]|(?R))*\}") - json_match = json_pattern.search(json_string) - - if json_match: - # Extract the valid JSON object from the string - json_string = json_match.group(0) - logger.typewriter_log( - title="Apparently json was fixed.", title_color=Fore.GREEN - ) - if CFG.speak_mode and CFG.debug_mode: - say_text("Apparently json was fixed.") - else: - raise ValueError("No valid JSON object found") - - except (json.JSONDecodeError, ValueError): - if CFG.debug_mode: - logger.error(f"Error: Invalid JSON: {json_string}\n") - if CFG.speak_mode: - say_text("Didn't work. I will have to ignore this response then.") - logger.error("Error: Invalid JSON, setting it to empty JSON now.\n") - json_string = {} - - return json_string - - -def balance_braces(json_string: str) -> str | None: -======= def balance_braces(json_string: str) -> Optional[str]: ->>>>>>> 67f32105 (Remove deprecated function from bracket_termination.py) """ Balance the braces in a JSON string. diff --git a/autogpt/json_fixes/parsing.py b/autogpt/json_fixes/parsing.py index d3a51f43..1e391eed 100644 --- a/autogpt/json_fixes/parsing.py +++ b/autogpt/json_fixes/parsing.py @@ -3,14 +3,9 @@ from __future__ import annotations import contextlib import json -<<<<<<< HEAD -from typing import Any - -======= from typing import Any, Dict, Union from colorama import Fore from regex import regex ->>>>>>> d3d8253b (Refactor parsing module and move JSON fix function to appropriate location) from autogpt.config import Config from autogpt.json_fixes.auto_fix import fix_json from autogpt.json_fixes.bracket_termination import balance_braces @@ -115,11 +110,7 @@ def fix_and_parse_json( def try_ai_fix( try_to_fix_with_gpt: bool, exception: Exception, json_to_load: str -<<<<<<< HEAD -) -> str | dict[Any, Any]: -======= ) -> Dict[Any, Any]: ->>>>>>> d3d8253b (Refactor parsing module and move JSON fix function to appropriate location) """Try to fix the JSON with the AI Args: From dc80a5a2ec6c7ceb2055894684ca7b680039a4c7 Mon Sep 17 00:00:00 2001 From: Jakub Bober Date: Sun, 16 Apr 2023 21:01:18 +0200 Subject: [PATCH 23/41] Add "Memory Backend Setup" subtitle Add the subtitle to match the Table of Contents --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index 21f3ccf2..7fce2e8f 100644 --- a/README.md +++ b/README.md @@ -280,6 +280,8 @@ To switch to either, change the `MEMORY_BACKEND` env variable to the value that * `milvus` will use the milvus cache that you configured * `weaviate` will use the weaviate cache that you configured +## Memory Backend Setup + ### Redis Setup > _**CAUTION**_ \ This is not intended to be publicly accessible and lacks security measures. Therefore, avoid exposing Redis to the internet without a password or at all From 627533bed631a15504b3584bf2aa70fe7b23aa86 Mon Sep 17 00:00:00 2001 From: 0xArty Date: Sun, 16 Apr 2023 21:55:53 +0100 Subject: [PATCH 24/41] minimall add pytest (#1859) * minimall add pytest * updated docs and pytest command * proveted milvus integration test running if milvus is not installed --- .pre-commit-config.yaml | 8 +- README.md | 19 +++- requirements.txt | 7 ++ tests/integration/milvus_memory_tests.py | 71 ++++++++------- tests/local_cache_test.py | 35 +++++--- tests/milvus_memory_test.py | 109 ++++++++++++----------- tests/smoke_test.py | 82 ++++++++--------- tests/unit/test_commands.py | 34 +++---- 8 files changed, 208 insertions(+), 157 deletions(-) diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index fb75cd59..dd1d0ec9 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -30,4 +30,10 @@ repos: language: python types: [ python ] exclude: .+/(dist|.venv|venv|build)/.+ - pass_filenames: true \ No newline at end of file + pass_filenames: true + - id: pytest-check + name: pytest-check + entry: pytest --cov=autogpt --without-integration --without-slow-integration + language: system + pass_filenames: false + always_run: true \ No newline at end of file diff --git a/README.md b/README.md index 58ed4d97..f60aa9ff 100644 --- a/README.md +++ b/README.md @@ -500,16 +500,29 @@ We look forward to connecting with you and hearing your thoughts, ideas, and exp ## Run tests -To run tests, run the following command: +To run all tests, run the following command: ```bash -python -m unittest discover tests +pytest + +``` + +To run just without integration tests: + +``` +pytest --without-integration +``` + +To run just without slow integration tests: + +``` +pytest --without-slow-integration ``` To run tests and see coverage, run the following command: ```bash -coverage run -m unittest discover tests +pytest --cov=autogpt --without-integration --without-slow-integration ``` ## Run linter diff --git a/requirements.txt b/requirements.txt index 64c2e4c0..843b66bf 100644 --- a/requirements.txt +++ b/requirements.txt @@ -29,5 +29,12 @@ black sourcery isort gitpython==3.1.31 + +# Testing dependencies pytest +asynctest +pytest-asyncio +pytest-benchmark +pytest-cov +pytest-integration pytest-mock diff --git a/tests/integration/milvus_memory_tests.py b/tests/integration/milvus_memory_tests.py index 96934cd6..ec38bf2f 100644 --- a/tests/integration/milvus_memory_tests.py +++ b/tests/integration/milvus_memory_tests.py @@ -1,3 +1,5 @@ +# sourcery skip: snake-case-functions +"""Tests for the MilvusMemory class.""" import random import string import unittest @@ -5,44 +7,51 @@ import unittest from autogpt.config import Config from autogpt.memory.milvus import MilvusMemory +try: -class TestMilvusMemory(unittest.TestCase): - def random_string(self, length): - return "".join(random.choice(string.ascii_letters) for _ in range(length)) + class TestMilvusMemory(unittest.TestCase): + """Tests for the MilvusMemory class.""" - def setUp(self): - cfg = Config() - cfg.milvus_addr = "localhost:19530" - self.memory = MilvusMemory(cfg) - self.memory.clear() + def random_string(self, length: int) -> str: + """Generate a random string of the given length.""" + return "".join(random.choice(string.ascii_letters) for _ in range(length)) - # Add example texts to the cache - self.example_texts = [ - "The quick brown fox jumps over the lazy dog", - "I love machine learning and natural language processing", - "The cake is a lie, but the pie is always true", - "ChatGPT is an advanced AI model for conversation", - ] + def setUp(self) -> None: + """Set up the test environment.""" + cfg = Config() + cfg.milvus_addr = "localhost:19530" + self.memory = MilvusMemory(cfg) + self.memory.clear() - for text in self.example_texts: - self.memory.add(text) + # Add example texts to the cache + self.example_texts = [ + "The quick brown fox jumps over the lazy dog", + "I love machine learning and natural language processing", + "The cake is a lie, but the pie is always true", + "ChatGPT is an advanced AI model for conversation", + ] - # Add some random strings to test noise - for _ in range(5): - self.memory.add(self.random_string(10)) + for text in self.example_texts: + self.memory.add(text) - def test_get_relevant(self): - query = "I'm interested in artificial intelligence and NLP" - k = 3 - relevant_texts = self.memory.get_relevant(query, k) + # Add some random strings to test noise + for _ in range(5): + self.memory.add(self.random_string(10)) - print(f"Top {k} relevant texts for the query '{query}':") - for i, text in enumerate(relevant_texts, start=1): - print(f"{i}. {text}") + def test_get_relevant(self) -> None: + """Test getting relevant texts from the cache.""" + query = "I'm interested in artificial intelligence and NLP" + num_relevant = 3 + relevant_texts = self.memory.get_relevant(query, num_relevant) - self.assertEqual(len(relevant_texts), k) - self.assertIn(self.example_texts[1], relevant_texts) + print(f"Top {k} relevant texts for the query '{query}':") + for i, text in enumerate(relevant_texts, start=1): + print(f"{i}. {text}") + self.assertEqual(len(relevant_texts), k) + self.assertIn(self.example_texts[1], relevant_texts) -if __name__ == "__main__": - unittest.main() +except: + print( + "Skipping tests/integration/milvus_memory_tests.py as Milvus is not installed." + ) diff --git a/tests/local_cache_test.py b/tests/local_cache_test.py index 91c922b0..fa596320 100644 --- a/tests/local_cache_test.py +++ b/tests/local_cache_test.py @@ -1,3 +1,5 @@ +# sourcery skip: snake-case-functions +"""Tests for LocalCache class""" import os import sys import unittest @@ -5,7 +7,8 @@ import unittest from autogpt.memory.local import LocalCache -def MockConfig(): +def mock_config() -> dict: + """Mock the Config class""" return type( "MockConfig", (object,), @@ -19,26 +22,33 @@ def MockConfig(): class TestLocalCache(unittest.TestCase): - def setUp(self): - self.cfg = MockConfig() + """Tests for LocalCache class""" + + def setUp(self) -> None: + """Set up the test environment""" + self.cfg = mock_config() self.cache = LocalCache(self.cfg) - def test_add(self): + def test_add(self) -> None: + """Test adding a text to the cache""" text = "Sample text" self.cache.add(text) self.assertIn(text, self.cache.data.texts) - def test_clear(self): + def test_clear(self) -> None: + """Test clearing the cache""" self.cache.clear() - self.assertEqual(self.cache.data, [""]) + self.assertEqual(self.cache.data.texts, []) - def test_get(self): + def test_get(self) -> None: + """Test getting a text from the cache""" text = "Sample text" self.cache.add(text) result = self.cache.get(text) self.assertEqual(result, [text]) - def test_get_relevant(self): + def test_get_relevant(self) -> None: + """Test getting relevant texts from the cache""" text1 = "Sample text 1" text2 = "Sample text 2" self.cache.add(text1) @@ -46,12 +56,9 @@ class TestLocalCache(unittest.TestCase): result = self.cache.get_relevant(text1, 1) self.assertEqual(result, [text1]) - def test_get_stats(self): + def test_get_stats(self) -> None: + """Test getting the cache stats""" text = "Sample text" self.cache.add(text) stats = self.cache.get_stats() - self.assertEqual(stats, (1, self.cache.data.embeddings.shape)) - - -if __name__ == "__main__": - unittest.main() + self.assertEqual(stats, (4, self.cache.data.embeddings.shape)) diff --git a/tests/milvus_memory_test.py b/tests/milvus_memory_test.py index 0113fa1c..e0e2f7fc 100644 --- a/tests/milvus_memory_test.py +++ b/tests/milvus_memory_test.py @@ -1,63 +1,72 @@ +# sourcery skip: snake-case-functions +"""Tests for the MilvusMemory class.""" import os import sys import unittest -from autogpt.memory.milvus import MilvusMemory +try: + from autogpt.memory.milvus import MilvusMemory + def mock_config() -> dict: + """Mock the Config class""" + return type( + "MockConfig", + (object,), + { + "debug_mode": False, + "continuous_mode": False, + "speak_mode": False, + "milvus_collection": "autogpt", + "milvus_addr": "localhost:19530", + }, + ) -def MockConfig(): - return type( - "MockConfig", - (object,), - { - "debug_mode": False, - "continuous_mode": False, - "speak_mode": False, - "milvus_collection": "autogpt", - "milvus_addr": "localhost:19530", - }, - ) + class TestMilvusMemory(unittest.TestCase): + """Tests for the MilvusMemory class.""" + def setUp(self) -> None: + """Set up the test environment""" + self.cfg = MockConfig() + self.memory = MilvusMemory(self.cfg) -class TestMilvusMemory(unittest.TestCase): - def setUp(self): - self.cfg = MockConfig() - self.memory = MilvusMemory(self.cfg) + def test_add(self) -> None: + """Test adding a text to the cache""" + text = "Sample text" + self.memory.clear() + self.memory.add(text) + result = self.memory.get(text) + self.assertEqual([text], result) - def test_add(self): - text = "Sample text" - self.memory.clear() - self.memory.add(text) - result = self.memory.get(text) - self.assertEqual([text], result) + def test_clear(self) -> None: + """Test clearing the cache""" + self.memory.clear() + self.assertEqual(self.memory.collection.num_entities, 0) - def test_clear(self): - self.memory.clear() - self.assertEqual(self.memory.collection.num_entities, 0) + def test_get(self) -> None: + """Test getting a text from the cache""" + text = "Sample text" + self.memory.clear() + self.memory.add(text) + result = self.memory.get(text) + self.assertEqual(result, [text]) - def test_get(self): - text = "Sample text" - self.memory.clear() - self.memory.add(text) - result = self.memory.get(text) - self.assertEqual(result, [text]) + def test_get_relevant(self) -> None: + """Test getting relevant texts from the cache""" + text1 = "Sample text 1" + text2 = "Sample text 2" + self.memory.clear() + self.memory.add(text1) + self.memory.add(text2) + result = self.memory.get_relevant(text1, 1) + self.assertEqual(result, [text1]) - def test_get_relevant(self): - text1 = "Sample text 1" - text2 = "Sample text 2" - self.memory.clear() - self.memory.add(text1) - self.memory.add(text2) - result = self.memory.get_relevant(text1, 1) - self.assertEqual(result, [text1]) + def test_get_stats(self) -> None: + """Test getting the cache stats""" + text = "Sample text" + self.memory.clear() + self.memory.add(text) + stats = self.memory.get_stats() + self.assertEqual(15, len(stats)) - def test_get_stats(self): - text = "Sample text" - self.memory.clear() - self.memory.add(text) - stats = self.memory.get_stats() - self.assertEqual(15, len(stats)) - - -if __name__ == "__main__": - unittest.main() +except: + print("Milvus not installed, skipping tests") diff --git a/tests/smoke_test.py b/tests/smoke_test.py index 50e97b7b..1b9d643f 100644 --- a/tests/smoke_test.py +++ b/tests/smoke_test.py @@ -1,31 +1,34 @@ +"""Smoke test for the autogpt package.""" import os import subprocess import sys -import unittest + +import pytest from autogpt.commands.file_operations import delete_file, read_file -env_vars = {"MEMORY_BACKEND": "no_memory", "TEMPERATURE": "0"} +@pytest.mark.integration_test +def test_write_file() -> None: + """ + Test case to check if the write_file command can successfully write 'Hello World' to a file + named 'hello_world.txt'. -class TestCommands(unittest.TestCase): - def test_write_file(self): - # Test case to check if the write_file command can successfully write 'Hello World' to a file - # named 'hello_world.txt'. + Read the current ai_settings.yaml file and store its content. + """ + env_vars = {"MEMORY_BACKEND": "no_memory", "TEMPERATURE": "0"} + ai_settings = None + if os.path.exists("ai_settings.yaml"): + with open("ai_settings.yaml", "r") as f: + ai_settings = f.read() + os.remove("ai_settings.yaml") - # Read the current ai_settings.yaml file and store its content. - ai_settings = None - if os.path.exists("ai_settings.yaml"): - with open("ai_settings.yaml", "r") as f: - ai_settings = f.read() - os.remove("ai_settings.yaml") - - try: - if os.path.exists("hello_world.txt"): - # Clean up any existing 'hello_world.txt' file before testing. - delete_file("hello_world.txt") - # Prepare input data for the test. - input_data = """write_file-GPT + try: + if os.path.exists("hello_world.txt"): + # Clean up any existing 'hello_world.txt' file before testing. + delete_file("hello_world.txt") + # Prepare input data for the test. + input_data = """write_file-GPT an AI designed to use the write_file command to write 'Hello World' into a file named "hello_world.txt" and then use the task_complete command to complete the task. Use the write_file command to write 'Hello World' into a file named "hello_world.txt". Use the task_complete command to complete the task. @@ -33,31 +36,24 @@ Do not use any other commands. y -5 EOF""" - command = f"{sys.executable} -m autogpt" + command = f"{sys.executable} -m autogpt" - # Execute the script with the input data. - process = subprocess.Popen( - command, - stdin=subprocess.PIPE, - shell=True, - env={**os.environ, **env_vars}, - ) - process.communicate(input_data.encode()) - - # Read the content of the 'hello_world.txt' file created during the test. - content = read_file("hello_world.txt") - finally: - if ai_settings: - # Restore the original ai_settings.yaml file. - with open("ai_settings.yaml", "w") as f: - f.write(ai_settings) - - # Check if the content of the 'hello_world.txt' file is equal to 'Hello World'. - self.assertEqual( - content, "Hello World", f"Expected 'Hello World', got {content}" + # Execute the script with the input data. + process = subprocess.Popen( + command, + stdin=subprocess.PIPE, + shell=True, + env={**os.environ, **env_vars}, ) + process.communicate(input_data.encode()) + # Read the content of the 'hello_world.txt' file created during the test. + content = read_file("hello_world.txt") + finally: + if ai_settings: + # Restore the original ai_settings.yaml file. + with open("ai_settings.yaml", "w") as f: + f.write(ai_settings) -# Run the test case. -if __name__ == "__main__": - unittest.main() + # Check if the content of the 'hello_world.txt' file is equal to 'Hello World'. + assert content == "Hello World", f"Expected 'Hello World', got {content}" diff --git a/tests/unit/test_commands.py b/tests/unit/test_commands.py index e15709aa..ecbac9b7 100644 --- a/tests/unit/test_commands.py +++ b/tests/unit/test_commands.py @@ -1,18 +1,22 @@ +"""Unit tests for the commands module""" +from unittest.mock import MagicMock, patch + +import pytest + import autogpt.agent.agent_manager as agent_manager -from autogpt.app import start_agent, list_agents, execute_command -import unittest -from unittest.mock import patch, MagicMock +from autogpt.app import execute_command, list_agents, start_agent -class TestCommands(unittest.TestCase): - def test_make_agent(self): - with patch("openai.ChatCompletion.create") as mock: - obj = MagicMock() - obj.response.choices[0].messages[0].content = "Test message" - mock.return_value = obj - start_agent("Test Agent", "chat", "Hello, how are you?", "gpt2") - agents = list_agents() - self.assertEqual("List of agents:\n0: chat", agents) - start_agent("Test Agent 2", "write", "Hello, how are you?", "gpt2") - agents = list_agents() - self.assertEqual("List of agents:\n0: chat\n1: write", agents) +@pytest.mark.integration_test +def test_make_agent() -> None: + """Test the make_agent command""" + with patch("openai.ChatCompletion.create") as mock: + obj = MagicMock() + obj.response.choices[0].messages[0].content = "Test message" + mock.return_value = obj + start_agent("Test Agent", "chat", "Hello, how are you?", "gpt2") + agents = list_agents() + assert "List of agents:\n0: chat" == agents + start_agent("Test Agent 2", "write", "Hello, how are you?", "gpt2") + agents = list_agents() + assert "List of agents:\n0: chat\n1: write" == agents From 147d3733bf068d8c71a901b8a0e31cfda5c4a687 Mon Sep 17 00:00:00 2001 From: 0xArty Date: Sun, 16 Apr 2023 16:03:22 +0100 Subject: [PATCH 25/41] Change ci to pytest --- .github/workflows/ci.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml index 366aaf67..39f3aea9 100644 --- a/.github/workflows/ci.yml +++ b/.github/workflows/ci.yml @@ -36,7 +36,7 @@ jobs: - name: Run unittest tests with coverage run: | - coverage run --source=autogpt -m unittest discover tests + pytest --cov=autogpt --without-integration --without-slow-integration - name: Generate coverage report run: | From 955a5b0a4357802a8142585ad78105f6342738ad Mon Sep 17 00:00:00 2001 From: 0xArty Date: Sun, 16 Apr 2023 16:13:16 +0100 Subject: [PATCH 26/41] Marked local chache tests as integration tests as they require api keys --- tests/local_cache_test.py | 3 +++ 1 file changed, 3 insertions(+) diff --git a/tests/local_cache_test.py b/tests/local_cache_test.py index fa596320..bb108626 100644 --- a/tests/local_cache_test.py +++ b/tests/local_cache_test.py @@ -4,6 +4,8 @@ import os import sys import unittest +import pytest + from autogpt.memory.local import LocalCache @@ -21,6 +23,7 @@ def mock_config() -> dict: ) +@pytest.mark.integration_test class TestLocalCache(unittest.TestCase): """Tests for LocalCache class""" From 8f0d553e4eaed9757f87ec33ec202cc7e570d8d5 Mon Sep 17 00:00:00 2001 From: Benedict Hobart Date: Sun, 16 Apr 2023 15:45:38 +0000 Subject: [PATCH 27/41] Improve dev containers so autogpt can browse the web --- .devcontainer/Dockerfile | 7 ++++++- .devcontainer/devcontainer.json | 1 + autogpt/commands/web_selenium.py | 1 + 3 files changed, 8 insertions(+), 1 deletion(-) diff --git a/.devcontainer/Dockerfile b/.devcontainer/Dockerfile index f3b2e2db..379f6310 100644 --- a/.devcontainer/Dockerfile +++ b/.devcontainer/Dockerfile @@ -1,6 +1,6 @@ # [Choice] Python version (use -bullseye variants on local arm64/Apple Silicon): 3, 3.10, 3.9, 3.8, 3.7, 3.6, 3-bullseye, 3.10-bullseye, 3.9-bullseye, 3.8-bullseye, 3.7-bullseye, 3.6-bullseye, 3-buster, 3.10-buster, 3.9-buster, 3.8-buster, 3.7-buster, 3.6-buster ARG VARIANT=3-bullseye -FROM python:3.8 +FROM --platform=linux/amd64 python:3.8 RUN apt-get update && export DEBIAN_FRONTEND=noninteractive \ # Remove imagemagick due to https://security-tracker.debian.org/tracker/CVE-2019-10131 @@ -10,6 +10,11 @@ RUN apt-get update && export DEBIAN_FRONTEND=noninteractive \ # They are installed by the base image (python) which does not have the patch. RUN python3 -m pip install --upgrade setuptools +# Install Chrome for web browsing +RUN apt-get update && export DEBIAN_FRONTEND=noninteractive \ + && curl -sSL https://dl.google.com/linux/direct/google-chrome-stable_current_$(dpkg --print-architecture).deb -o /tmp/chrome.deb \ + && apt-get -y install /tmp/chrome.deb + # [Optional] If your pip requirements rarely change, uncomment this section to add them to the image. # COPY requirements.txt /tmp/pip-tmp/ # RUN pip3 --disable-pip-version-check --no-cache-dir install -r /tmp/pip-tmp/requirements.txt \ diff --git a/.devcontainer/devcontainer.json b/.devcontainer/devcontainer.json index 5fefd9c1..f26810fb 100644 --- a/.devcontainer/devcontainer.json +++ b/.devcontainer/devcontainer.json @@ -11,6 +11,7 @@ "userGid": "1000", "upgradePackages": "true" }, + "ghcr.io/devcontainers/features/desktop-lite:1": {}, "ghcr.io/devcontainers/features/python:1": "none", "ghcr.io/devcontainers/features/node:1": "none", "ghcr.io/devcontainers/features/git:1": { diff --git a/autogpt/commands/web_selenium.py b/autogpt/commands/web_selenium.py index 1d078d76..8c652294 100644 --- a/autogpt/commands/web_selenium.py +++ b/autogpt/commands/web_selenium.py @@ -75,6 +75,7 @@ def scrape_text_with_selenium(url: str) -> tuple[WebDriver, str]: # See https://developer.apple.com/documentation/webkit/testing_with_webdriver_in_safari driver = webdriver.Safari(options=options) else: + options.add_argument("--no-sandbox") driver = webdriver.Chrome( executable_path=ChromeDriverManager().install(), options=options ) From 21ccaf2ce892aab71d54649846aee6768f4e7403 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 14:16:48 -0700 Subject: [PATCH 28/41] Refactor variable names and remove unnecessary blank lines in __main__.py --- autogpt/__main__.py | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) diff --git a/autogpt/__main__.py b/autogpt/__main__.py index 29ccddbf..7fe6aec3 100644 --- a/autogpt/__main__.py +++ b/autogpt/__main__.py @@ -3,13 +3,10 @@ import logging from colorama import Fore from autogpt.agent.agent import Agent from autogpt.args import parse_arguments - from autogpt.config import Config, check_openai_api_key from autogpt.logs import logger from autogpt.memory import get_memory - from autogpt.prompt import construct_prompt - # Load environment variables from .env file @@ -21,13 +18,13 @@ def main() -> None: parse_arguments() logger.set_level(logging.DEBUG if cfg.debug_mode else logging.INFO) ai_name = "" - prompt = construct_prompt() + master_prompt = construct_prompt() # print(prompt) # Initialize variables full_message_history = [] next_action_count = 0 # Make a constant: - user_input = ( + triggering_prompt = ( "Determine which next command to use, and respond using the" " format specified above:" ) @@ -43,8 +40,8 @@ def main() -> None: memory=memory, full_message_history=full_message_history, next_action_count=next_action_count, - prompt=prompt, - user_input=user_input, + master_prompt=master_prompt, + triggering_prompt=triggering_prompt, ) agent.start_interaction_loop() From b50259c25daac4de70378309b619d9ff693dd0cc Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 14:16:48 -0700 Subject: [PATCH 29/41] Update variable names, improve comments, and modify input handling in agent.py --- autogpt/agent/agent.py | 43 +++++++++++++++++++++++++----------------- 1 file changed, 26 insertions(+), 17 deletions(-) diff --git a/autogpt/agent/agent.py b/autogpt/agent/agent.py index 32d982e5..3be17a89 100644 --- a/autogpt/agent/agent.py +++ b/autogpt/agent/agent.py @@ -19,9 +19,18 @@ class Agent: memory: The memory object to use. full_message_history: The full message history. next_action_count: The number of actions to execute. - prompt: The prompt to use. - user_input: The user input. + master_prompt: The master prompt is the initial prompt that defines everything the AI needs to know to achieve its task successfully. + Currently, the dynamic and customizable information in the master prompt are ai_name, description and goals. + triggering_prompt: The last sentence the AI will see before answering. For Auto-GPT, this prompt is: + Determine which next command to use, and respond using the format specified above: + The triggering prompt is not part of the master prompt because between the master prompt and the triggering + prompt we have contextual information that can distract the AI and make it forget that its goal is to find the next task to achieve. + MASTER PROMPT + CONTEXTUAL INFORMATION (memory, previous conversations, anything relevant) + TRIGGERING PROMPT + + The triggering prompt reminds the AI about its short term meta task (defining the next task) """ def __init__( @@ -30,15 +39,15 @@ class Agent: memory, full_message_history, next_action_count, - prompt, - user_input, + master_prompt, + triggering_prompt, ): self.ai_name = ai_name self.memory = memory self.full_message_history = full_message_history self.next_action_count = next_action_count - self.prompt = prompt - self.user_input = user_input + self.master_prompt = master_prompt + self.triggering_prompt = triggering_prompt def start_interaction_loop(self): # Interaction Loop @@ -62,8 +71,8 @@ class Agent: # Send message to AI, get response with Spinner("Thinking... "): assistant_reply = chat_with_ai( - self.prompt, - self.user_input, + self.master_prompt, + self.triggering_prompt, self.full_message_history, self.memory, cfg.fast_token_limit, @@ -88,7 +97,7 @@ class Agent: ### GET USER AUTHORIZATION TO EXECUTE COMMAND ### # Get key press: Prompt the user to press enter to continue or escape # to exit - self.user_input = "" + user_input = "" logger.typewriter_log( "NEXT ACTION: ", Fore.CYAN, @@ -106,14 +115,14 @@ class Agent: Fore.MAGENTA + "Input:" + Style.RESET_ALL ) if console_input.lower().rstrip() == "y": - self.user_input = "GENERATE NEXT COMMAND JSON" + user_input = "GENERATE NEXT COMMAND JSON" break elif console_input.lower().startswith("y -"): try: self.next_action_count = abs( int(console_input.split(" ")[1]) ) - self.user_input = "GENERATE NEXT COMMAND JSON" + user_input = "GENERATE NEXT COMMAND JSON" except ValueError: print( "Invalid input format. Please enter 'y -n' where n is" @@ -122,20 +131,20 @@ class Agent: continue break elif console_input.lower() == "n": - self.user_input = "EXIT" + user_input = "EXIT" break else: - self.user_input = console_input + user_input = console_input command_name = "human_feedback" break - if self.user_input == "GENERATE NEXT COMMAND JSON": + if user_input == "GENERATE NEXT COMMAND JSON": logger.typewriter_log( "-=-=-=-=-=-=-= COMMAND AUTHORISED BY USER -=-=-=-=-=-=-=", Fore.MAGENTA, "", ) - elif self.user_input == "EXIT": + elif user_input == "EXIT": print("Exiting...", flush=True) break else: @@ -153,7 +162,7 @@ class Agent: f"Command {command_name} threw the following error: {arguments}" ) elif command_name == "human_feedback": - result = f"Human feedback: {self.user_input}" + result = f"Human feedback: {user_input}" else: result = ( f"Command {command_name} returned: " @@ -165,7 +174,7 @@ class Agent: memory_to_add = ( f"Assistant Reply: {assistant_reply} " f"\nResult: {result} " - f"\nHuman Feedback: {self.user_input} " + f"\nHuman Feedback: {user_input} " ) self.memory.add(memory_to_add) From b5e0127b16bb88f6b6e18ada0efabc1422c9f3de Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 14:16:48 -0700 Subject: [PATCH 30/41] Only print JSON object validation message in debug mode --- autogpt/json_validation/validate_json.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/autogpt/json_validation/validate_json.py b/autogpt/json_validation/validate_json.py index 127fcc17..440c3b0b 100644 --- a/autogpt/json_validation/validate_json.py +++ b/autogpt/json_validation/validate_json.py @@ -24,7 +24,7 @@ def validate_json(json_object: object, schema_name: object) -> object: for error in errors: logger.error(f"Error: {error.message}") - else: + elif CFG.debug_mode: print("The JSON object is valid.") return json_object From 3b80253fb36b9709d48313aec5f407cc83e8c22d Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 14:16:48 -0700 Subject: [PATCH 31/41] Update process creation in benchmark script --- benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py b/benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py index d6cae972..f7f1dac9 100644 --- a/benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py +++ b/benchmark/benchmark_entrepeneur_gpt_with_difficult_user.py @@ -73,9 +73,12 @@ Disappointing suggestion. Not helpful. Needs improvement. Not what I need.''' + # TODO: add questions above, to distract it even more. + command = f'{sys.executable} -m autogpt' - process = subprocess.Popen(command, stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True) + process = subprocess.Popen(command, stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE, + shell=True) stdout_output, stderr_output = process.communicate(input_data.encode()) From 89e0e8992795accfc41183723064dcdab9719f8e Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 14:22:58 -0700 Subject: [PATCH 32/41] change master prompt to system prompt --- autogpt/__main__.py | 4 ++-- autogpt/agent/agent.py | 14 +++++++------- 2 files changed, 9 insertions(+), 9 deletions(-) diff --git a/autogpt/__main__.py b/autogpt/__main__.py index 7fe6aec3..5f462234 100644 --- a/autogpt/__main__.py +++ b/autogpt/__main__.py @@ -18,7 +18,7 @@ def main() -> None: parse_arguments() logger.set_level(logging.DEBUG if cfg.debug_mode else logging.INFO) ai_name = "" - master_prompt = construct_prompt() + system_prompt = construct_prompt() # print(prompt) # Initialize variables full_message_history = [] @@ -40,7 +40,7 @@ def main() -> None: memory=memory, full_message_history=full_message_history, next_action_count=next_action_count, - master_prompt=master_prompt, + system_prompt=system_prompt, triggering_prompt=triggering_prompt, ) agent.start_interaction_loop() diff --git a/autogpt/agent/agent.py b/autogpt/agent/agent.py index 3be17a89..9853f6a0 100644 --- a/autogpt/agent/agent.py +++ b/autogpt/agent/agent.py @@ -19,14 +19,14 @@ class Agent: memory: The memory object to use. full_message_history: The full message history. next_action_count: The number of actions to execute. - master_prompt: The master prompt is the initial prompt that defines everything the AI needs to know to achieve its task successfully. - Currently, the dynamic and customizable information in the master prompt are ai_name, description and goals. + system_prompt: The system prompt is the initial prompt that defines everything the AI needs to know to achieve its task successfully. + Currently, the dynamic and customizable information in the system prompt are ai_name, description and goals. triggering_prompt: The last sentence the AI will see before answering. For Auto-GPT, this prompt is: Determine which next command to use, and respond using the format specified above: - The triggering prompt is not part of the master prompt because between the master prompt and the triggering + The triggering prompt is not part of the system prompt because between the system prompt and the triggering prompt we have contextual information that can distract the AI and make it forget that its goal is to find the next task to achieve. - MASTER PROMPT + SYSTEM PROMPT CONTEXTUAL INFORMATION (memory, previous conversations, anything relevant) TRIGGERING PROMPT @@ -39,14 +39,14 @@ class Agent: memory, full_message_history, next_action_count, - master_prompt, + system_prompt, triggering_prompt, ): self.ai_name = ai_name self.memory = memory self.full_message_history = full_message_history self.next_action_count = next_action_count - self.master_prompt = master_prompt + self.system_prompt = system_prompt self.triggering_prompt = triggering_prompt def start_interaction_loop(self): @@ -71,7 +71,7 @@ class Agent: # Send message to AI, get response with Spinner("Thinking... "): assistant_reply = chat_with_ai( - self.master_prompt, + self.system_prompt, self.triggering_prompt, self.full_message_history, self.memory, From 4f33e1bf89e580355dfcf6890779799c584e9563 Mon Sep 17 00:00:00 2001 From: k-boikov Date: Sun, 16 Apr 2023 18:38:08 +0300 Subject: [PATCH 33/41] add utf-8 encoding to file handlers for logging --- autogpt/logs.py | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/autogpt/logs.py b/autogpt/logs.py index f18e2140..c1e436db 100644 --- a/autogpt/logs.py +++ b/autogpt/logs.py @@ -46,7 +46,9 @@ class Logger(metaclass=Singleton): self.console_handler.setFormatter(console_formatter) # Info handler in activity.log - self.file_handler = logging.FileHandler(os.path.join(log_dir, log_file)) + self.file_handler = logging.FileHandler( + os.path.join(log_dir, log_file), 'a', 'utf-8' + ) self.file_handler.setLevel(logging.DEBUG) info_formatter = AutoGptFormatter( "%(asctime)s %(levelname)s %(title)s %(message_no_color)s" @@ -54,7 +56,9 @@ class Logger(metaclass=Singleton): self.file_handler.setFormatter(info_formatter) # Error handler error.log - error_handler = logging.FileHandler(os.path.join(log_dir, error_file)) + error_handler = logging.FileHandler( + os.path.join(log_dir, error_file), 'a', 'utf-8' + ) error_handler.setLevel(logging.ERROR) error_formatter = AutoGptFormatter( "%(asctime)s %(levelname)s %(module)s:%(funcName)s:%(lineno)d %(title)s" From 1513be4acdcc85b27869219938ed90610a7db673 Mon Sep 17 00:00:00 2001 From: Merwane Hamadi Date: Sun, 16 Apr 2023 15:31:53 -0700 Subject: [PATCH 34/41] hotfix user input --- autogpt/agent/agent.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/autogpt/agent/agent.py b/autogpt/agent/agent.py index 9853f6a0..dca614c7 100644 --- a/autogpt/agent/agent.py +++ b/autogpt/agent/agent.py @@ -55,6 +55,8 @@ class Agent: loop_count = 0 command_name = None arguments = None + user_input = "" + while True: # Discontinue if continuous limit is reached loop_count += 1 @@ -97,7 +99,6 @@ class Agent: ### GET USER AUTHORIZATION TO EXECUTE COMMAND ### # Get key press: Prompt the user to press enter to continue or escape # to exit - user_input = "" logger.typewriter_log( "NEXT ACTION: ", Fore.CYAN, From c71c61dc584a41d72e2b27b02fe75a9f64e3e029 Mon Sep 17 00:00:00 2001 From: Adrian Scott Date: Sun, 16 Apr 2023 18:14:16 -0500 Subject: [PATCH 35/41] Added one space after period for better formatting --- autogpt/memory/local.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/autogpt/memory/local.py b/autogpt/memory/local.py index 6c7ee1b3..9b911eef 100644 --- a/autogpt/memory/local.py +++ b/autogpt/memory/local.py @@ -54,7 +54,7 @@ class LocalCache(MemoryProviderSingleton): self.data = CacheContent() else: print( - f"Warning: The file '{self.filename}' does not exist." + f"Warning: The file '{self.filename}' does not exist. " "Local memory would not be saved to a file." ) self.data = CacheContent() From 15059c2090be47d2a674113f509618b3f58a3510 Mon Sep 17 00:00:00 2001 From: Chris Cheney Date: Sun, 16 Apr 2023 17:28:25 -0500 Subject: [PATCH 36/41] ensure git operations occur in the working directory --- autogpt/commands/git_operations.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/autogpt/commands/git_operations.py b/autogpt/commands/git_operations.py index 3ff35cf3..675eb228 100644 --- a/autogpt/commands/git_operations.py +++ b/autogpt/commands/git_operations.py @@ -1,6 +1,7 @@ """Git operations for autogpt""" import git from autogpt.config import Config +from autogpt.workspace import path_in_workspace CFG = Config() @@ -16,8 +17,9 @@ def clone_repository(repo_url: str, clone_path: str) -> str: str: The result of the clone operation""" split_url = repo_url.split("//") auth_repo_url = f"//{CFG.github_username}:{CFG.github_api_key}@".join(split_url) + safe_clone_path = path_in_workspace(clone_path) try: - git.Repo.clone_from(auth_repo_url, clone_path) - return f"""Cloned {repo_url} to {clone_path}""" + git.Repo.clone_from(auth_repo_url, safe_clone_path) + return f"""Cloned {repo_url} to {safe_clone_path}""" except Exception as e: return f"Error: {str(e)}" From 56ecbeeef734019d1d3112fc2a657a14bff69ccb Mon Sep 17 00:00:00 2001 From: Reinier van der Leer Date: Mon, 17 Apr 2023 02:22:18 +0200 Subject: [PATCH 37/41] Clean up README --- README.md | 227 ++++++++++++++++++++++++------------------------------ 1 file changed, 102 insertions(+), 125 deletions(-) diff --git a/README.md b/README.md index f60aa9ff..07a088e3 100644 --- a/README.md +++ b/README.md @@ -1,15 +1,19 @@ # Auto-GPT: An Autonomous GPT-4 Experiment +[![Unit Tests](https://img.shields.io/github/actions/workflow/status/Significant-Gravitas/Auto-GPT/ci.yml?label=unit%20tests)](https://github.com/Significant-Gravitas/Auto-GPT/actions/workflows/ci.yml) +[![Discord Follow](https://dcbadge.vercel.app/api/server/autogpt?style=flat)](https://discord.gg/autogpt) +[![GitHub Repo stars](https://img.shields.io/github/stars/Significant-Gravitas/auto-gpt?style=social)](https://github.com/Significant-Gravitas/Auto-GPT/stargazers) +[![Twitter Follow](https://img.shields.io/twitter/follow/siggravitas?style=social)](https://twitter.com/SigGravitas) + + +
+ ### πŸ”΄ πŸ”΄ πŸ”΄ Urgent: USE `stable` not `master` πŸ”΄ πŸ”΄ πŸ”΄ -> This information takes precedence over any other information provided and should be followed accordingly. +**Download the latest `stable` release from here: https://github.com/Significant-Gravitas/Auto-GPT/releases/latest.** +The `master` branch may often be in a **broken** state. -Our workflow has been improved, but please note that `master` branch may often be in a **broken** state. -Please download the latest `stable` release from here: https://github.com/Torantulino/Auto-GPT/releases/latest. +
-![GitHub Repo stars](https://img.shields.io/github/stars/Significant-Gravitas/auto-gpt?style=social) -[![Twitter Follow](https://img.shields.io/twitter/follow/siggravitas?style=social)](https://twitter.com/SigGravitas) -[![Discord Follow](https://dcbadge.vercel.app/api/server/autogpt?style=flat)](https://discord.gg/autogpt) -[![Unit Tests](https://github.com/Significant-Gravitaso/Auto-GPT/actions/workflows/ci.yml/badge.svg)](https://github.com/Significant-Gravitas/Auto-GPT/actions/workflows/ci.yml) Auto-GPT is an experimental open-source application showcasing the capabilities of the GPT-4 language model. This program, driven by GPT-4, chains together LLM "thoughts", to autonomously achieve whatever goal you set. As one of the first examples of GPT-4 running fully autonomously, Auto-GPT pushes the boundaries of what is possible with AI. @@ -37,42 +41,6 @@ Development of this free, open-source project is made possible by all the Dradstone  CrypteorCapital  avy-ai  shawnharmsen  sunchongren  DailyBotHQ  mathewhawkins  MediConCenHK  kMag410  nicoguyon  Mobivs  jazgarewal  marv-technology  rapidstartup  Brodie0  lucas-chu  rejunity  comet-ml  ColinConwell  cfarquhar  ikarosai  ChrisDMT  Odin519Tomas  vkozacek  belharethsami  sultanmeghji  scryptedinc  johnculkin  RealChrisSean  fruition  jd3655  Web3Capital  allenstecat  tob-le-rone  SwftCoins  MetaPath01  joaomdmoura  ternary5  refinery1  josephcmiller2  webbcolton  tommygeee  lmaugustin  garythebat  Cameron-Fulton  angiaou  caitlynmeeks  MBassi91  Daniel1357  omphos  abhinav-pandey29  DataMetis  concreit  st617  RThaweewat  KiaArmani  Pythagora-io  AryaXAI  fabrietech  jun784  Mr-Bishop42  rickscode  projectonegames  rocks6  GalaxyVideoAgency  thisisjeffchen  TheStoneMX  txtr99  ZERO-A-ONE  

- - -## Table of Contents - -- [Auto-GPT: An Autonomous GPT-4 Experiment](#auto-gpt-an-autonomous-gpt-4-experiment) - - [πŸ”΄ πŸ”΄ πŸ”΄ Urgent: USE `stable` not `master` πŸ”΄ πŸ”΄ πŸ”΄](#----urgent-use-stable-not-master----) - - [Demo (30/03/2023):](#demo-30032023) - - [Table of Contents](#table-of-contents) - - [πŸš€ Features](#-features) - - [πŸ“‹ Requirements](#-requirements) - - [πŸ’Ύ Installation](#-installation) - - [πŸ”§ Usage](#-usage) - - [Logs](#logs) - - [Docker](#docker) - - [Command Line Arguments](#command-line-arguments) - - [πŸ—£οΈ Speech Mode](#️-speech-mode) - - [πŸ” Google API Keys Configuration](#-google-api-keys-configuration) - - [Setting up environment variables](#setting-up-environment-variables) - - [Memory Backend Setup](#memory-backend-setup) - - [Redis Setup](#redis-setup) - - [🌲 Pinecone API Key Setup](#-pinecone-api-key-setup) - - [Milvus Setup](#milvus-setup) - - [Weaviate Setup](#weaviate-setup) - - [Setting up environment variables](#setting-up-environment-variables-1) - - [Setting Your Cache Type](#setting-your-cache-type) - - [View Memory Usage](#view-memory-usage) - - [🧠 Memory pre-seeding](#-memory-pre-seeding) - - [πŸ’€ Continuous Mode ⚠️](#-continuous-mode-️) - - [GPT3.5 ONLY Mode](#gpt35-only-mode) - - [πŸ–Ό Image Generation](#-image-generation) - - [⚠️ Limitations](#️-limitations) - - [πŸ›‘ Disclaimer](#-disclaimer) - - [🐦 Connect with Us on Twitter](#-connect-with-us-on-twitter) - - [Run tests](#run-tests) - - [Run linter](#run-linter) - ## πŸš€ Features - 🌐 Internet access for searches and information gathering @@ -83,16 +51,17 @@ Development of this free, open-source project is made possible by all the

Blake Werlinger +

πŸ’– Help Fund Auto-GPT's Development πŸ’–

If you can spare a coffee, you can help to cover the costs of developing Auto-GPT and help push the boundaries of fully autonomous AI! From 9589334a305198c837bfb8720ed6f06176b2f216 Mon Sep 17 00:00:00 2001 From: EH Date: Mon, 17 Apr 2023 03:34:02 +0100 Subject: [PATCH 41/41] Add File Downloading Capabilities (#1680) * Added 'download_file' command * Added util and fixed spinner * Fixed comma and added autogpt/auto_gpt_workspace to .gitignore * Fix linter issues * Fix more linter issues * Fix Lint Issues * Added 'download_file' command * Added util and fixed spinner * Fixed comma and added autogpt/auto_gpt_workspace to .gitignore * Fix linter issues * Fix more linter issues * Conditionally add the 'download_file' prompt * Update args.py * Removed Duplicate Prompt * Switched to using path_in_workspace function --- .gitignore | 1 + autogpt/app.py | 5 +++ autogpt/args.py | 16 +++++++++- autogpt/commands/file_operations.py | 49 ++++++++++++++++++++++++++++- autogpt/config/config.py | 1 + autogpt/prompt.py | 10 ++++++ autogpt/spinner.py | 15 ++++++++- autogpt/utils.py | 13 ++++++++ 8 files changed, 107 insertions(+), 3 deletions(-) diff --git a/.gitignore b/.gitignore index eda7f327..2220ef6e 100644 --- a/.gitignore +++ b/.gitignore @@ -3,6 +3,7 @@ autogpt/keys.py autogpt/*json autogpt/node_modules/ autogpt/__pycache__/keys.cpython-310.pyc +autogpt/auto_gpt_workspace package-lock.json *.pyc auto_gpt_workspace/* diff --git a/autogpt/app.py b/autogpt/app.py index 78b5bd2f..19c075f0 100644 --- a/autogpt/app.py +++ b/autogpt/app.py @@ -17,6 +17,7 @@ from autogpt.commands.file_operations import ( read_file, search_files, write_to_file, + download_file ) from autogpt.json_fixes.parsing import fix_and_parse_json from autogpt.memory import get_memory @@ -164,6 +165,10 @@ def execute_command(command_name: str, arguments): return delete_file(arguments["file"]) elif command_name == "search_files": return search_files(arguments["directory"]) + elif command_name == "download_file": + if not CFG.allow_downloads: + return "Error: You do not have user authorization to download files locally." + return download_file(arguments["url"], arguments["file"]) elif command_name == "browse_website": return browse_website(arguments["url"], arguments["question"]) # TODO: Change these to take in a file rather than pasted code, if diff --git a/autogpt/args.py b/autogpt/args.py index eca32334..f0e9c07a 100644 --- a/autogpt/args.py +++ b/autogpt/args.py @@ -1,7 +1,7 @@ """This module contains the argument parsing logic for the script.""" import argparse -from colorama import Fore +from colorama import Fore, Back, Style from autogpt import utils from autogpt.config import Config from autogpt.logs import logger @@ -63,6 +63,12 @@ def parse_arguments() -> None: help="Specifies which ai_settings.yaml file to use, will also automatically" " skip the re-prompt.", ) + parser.add_argument( + '--allow-downloads', + action='store_true', + dest='allow_downloads', + help='Dangerous: Allows Auto-GPT to download files natively.' + ) args = parser.parse_args() if args.debug: @@ -133,5 +139,13 @@ def parse_arguments() -> None: CFG.ai_settings_file = file CFG.skip_reprompt = True + if args.allow_downloads: + logger.typewriter_log("Native Downloading:", Fore.GREEN, "ENABLED") + logger.typewriter_log("WARNING: ", Fore.YELLOW, + f"{Back.LIGHTYELLOW_EX}Auto-GPT will now be able to download and save files to your machine.{Back.RESET} " + + "It is recommended that you monitor any files it downloads carefully.") + logger.typewriter_log("WARNING: ", Fore.YELLOW, f"{Back.RED + Style.BRIGHT}ALWAYS REMEMBER TO NEVER OPEN FILES YOU AREN'T SURE OF!{Style.RESET_ALL}") + CFG.allow_downloads = True + if args.browser_name: CFG.selenium_web_browser = args.browser_name diff --git a/autogpt/commands/file_operations.py b/autogpt/commands/file_operations.py index 8abc2e23..d273c1a3 100644 --- a/autogpt/commands/file_operations.py +++ b/autogpt/commands/file_operations.py @@ -4,9 +4,16 @@ from __future__ import annotations import os import os.path from pathlib import Path -from typing import Generator +from typing import Generator, List +import requests +from requests.adapters import HTTPAdapter +from requests.adapters import Retry +from colorama import Fore, Back +from autogpt.spinner import Spinner +from autogpt.utils import readable_file_size from autogpt.workspace import path_in_workspace, WORKSPACE_PATH + LOG_FILE = "file_logger.txt" LOG_FILE_PATH = WORKSPACE_PATH / LOG_FILE @@ -214,3 +221,43 @@ def search_files(directory: str) -> list[str]: found_files.append(relative_path) return found_files + + +def download_file(url, filename): + """Downloads a file + Args: + url (str): URL of the file to download + filename (str): Filename to save the file as + """ + safe_filename = path_in_workspace(filename) + try: + message = f"{Fore.YELLOW}Downloading file from {Back.LIGHTBLUE_EX}{url}{Back.RESET}{Fore.RESET}" + with Spinner(message) as spinner: + session = requests.Session() + retry = Retry(total=3, backoff_factor=1, status_forcelist=[502, 503, 504]) + adapter = HTTPAdapter(max_retries=retry) + session.mount('http://', adapter) + session.mount('https://', adapter) + + total_size = 0 + downloaded_size = 0 + + with session.get(url, allow_redirects=True, stream=True) as r: + r.raise_for_status() + total_size = int(r.headers.get('Content-Length', 0)) + downloaded_size = 0 + + with open(safe_filename, 'wb') as f: + for chunk in r.iter_content(chunk_size=8192): + f.write(chunk) + downloaded_size += len(chunk) + + # Update the progress message + progress = f"{readable_file_size(downloaded_size)} / {readable_file_size(total_size)}" + spinner.update_message(f"{message} {progress}") + + return f'Successfully downloaded and locally stored file: "{filename}"! (Size: {readable_file_size(total_size)})' + except requests.HTTPError as e: + return f"Got an HTTP Error whilst trying to download file: {e}" + except Exception as e: + return "Error: " + str(e) diff --git a/autogpt/config/config.py b/autogpt/config/config.py index 22da52b0..fe6f4f32 100644 --- a/autogpt/config/config.py +++ b/autogpt/config/config.py @@ -24,6 +24,7 @@ class Config(metaclass=Singleton): self.continuous_limit = 0 self.speak_mode = False self.skip_reprompt = False + self.allow_downloads = False self.selenium_web_browser = os.getenv("USE_WEB_BROWSER", "chrome") self.ai_settings_file = os.getenv("AI_SETTINGS_FILE", "ai_settings.yaml") diff --git a/autogpt/prompt.py b/autogpt/prompt.py index 18a5736c..a2b20b1f 100644 --- a/autogpt/prompt.py +++ b/autogpt/prompt.py @@ -105,6 +105,16 @@ def get_prompt() -> str: ), ) + # Only add the download file command if the AI is allowed to execute it + if cfg.allow_downloads: + commands.append( + ( + "Downloads a file from the internet, and stores it locally", + "download_file", + {"url": "", "file": ""} + ), + ) + # Add these command last. commands.append( ("Do Nothing", "do_nothing", {}), diff --git a/autogpt/spinner.py b/autogpt/spinner.py index 56b4f20a..febcea8e 100644 --- a/autogpt/spinner.py +++ b/autogpt/spinner.py @@ -29,12 +29,14 @@ class Spinner: time.sleep(self.delay) sys.stdout.write(f"\r{' ' * (len(self.message) + 2)}\r") - def __enter__(self) -> None: + def __enter__(self): """Start the spinner""" self.running = True self.spinner_thread = threading.Thread(target=self.spin) self.spinner_thread.start() + return self + def __exit__(self, exc_type, exc_value, exc_traceback) -> None: """Stop the spinner @@ -48,3 +50,14 @@ class Spinner: self.spinner_thread.join() sys.stdout.write(f"\r{' ' * (len(self.message) + 2)}\r") sys.stdout.flush() + + def update_message(self, new_message, delay=0.1): + """Update the spinner message + Args: + new_message (str): New message to display + delay: Delay in seconds before updating the message + """ + time.sleep(delay) + sys.stdout.write(f"\r{' ' * (len(self.message) + 2)}\r") # Clear the current message + sys.stdout.flush() + self.message = new_message diff --git a/autogpt/utils.py b/autogpt/utils.py index 59709d02..11d98d1b 100644 --- a/autogpt/utils.py +++ b/autogpt/utils.py @@ -24,3 +24,16 @@ def validate_yaml_file(file: str): ) return (True, f"Successfully validated {Fore.CYAN}`{file}`{Fore.RESET}!") + + +def readable_file_size(size, decimal_places=2): + """Converts the given size in bytes to a readable format. + Args: + size: Size in bytes + decimal_places (int): Number of decimal places to display + """ + for unit in ['B', 'KB', 'MB', 'GB', 'TB']: + if size < 1024.0: + break + size /= 1024.0 + return f"{size:.{decimal_places}f} {unit}"