aljaz/Auto-GPT
1
0
Fork 0
mirror of https://github.com/aljazceru/Auto-GPT.git synced 2025-12-17 22:14:28 +01:00
Code Issues Packages Projects Releases Wiki Activity
243 Commits 1 Branch 0 Tags
8a08b0c01b5be1b70e58a1d6817d4f85c54465ef
Commit Graph

68 Commits

Author SHA1 Message Date
Preston Jensen
8a08b0c01b ask questions on websites 2023-04-03 19:20:42 -06:00
Toran Bruce Richards
f6c201f038 Merge pull request #102 from yousefissa/yi-fix-print-thoughts-failures 
fix: UnboundLocalError: local variable 'assistant_thoughts_reasoning' referenced before assignment
 2023-04-04 11:38:35 +12:00
Toran Bruce Richards
e612b119d3 Merge pull request #81 from yousefissa/yi-write-to-file 
Fix: create file dir if it doesnt exist during write_to_file
 2023-04-04 11:27:37 +12:00
Toran Bruce Richards
7e529e19d9 Removes print. 2023-04-04 00:24:22 +01:00
yousefissa
8753eba22c comments 2023-04-03 14:55:30 -07:00
yousefissa
04c43432c2 cleanup method 2023-04-03 14:53:19 -07:00
russellocean
f2ba7f21c5 Added support for Google Custom Search API 
This pull request adds functionality to the project to allow for a choice between the original Google search method and the Google Custom Search API. The google_search method uses the original method of scraping the HTML from the search results page, using googlesearch-python, while the google_official_search method uses the Google Custom Search API to retrieve search results.

How to test:

To test the functionality, ensure that you have valid API keys and search engine IDs for both the Google search method and the Google Custom Search API. You can set these values in your environment variables as described in the README.md file.

Additional Notes:

This pull request only adds functionality and makes improvements to existing code. No new features or major changes have been introduced.
 2023-04-03 16:44:10 -04:00
yousefissa
8a5c9800e7 fix assistant plan variables being referenced before assignment 2023-04-03 13:00:50 -07:00
yousefissa
9ef4fab084 error handling back 2023-04-03 08:50:07 -07:00
yousefissa
4416aa1aa1 create file dir if it doesnt exist during write_to_file 2023-04-03 08:48:43 -07:00
Toran Bruce Richards
51e293f64c Adds --gpt3only mode! 2023-04-03 15:46:46 +01:00
0xcha05
c8149b0415 better arg parsing 2023-04-03 19:56:01 +05:30
Veylkh
f72afc7558 dirtyjson -> json 
Fixes the missing UserDict (caused by migration from python 3.9 to ^3.10)
 2023-04-03 14:28:22 +02:00
Toran Bruce Richards
3093ec9369 Tweaks prompt for better results on both models! 2023-04-03 11:37:43 +01:00
Toran Bruce Richards
109b9288a7 Removes logging and uses new config token limit. 2023-04-03 11:37:18 +01:00
Toran Bruce Richards
e50e826857 Tweaks welcome back prompt. 2023-04-03 11:35:08 +01:00
Toran Bruce Richards
f426b51038 Changes python version to 3.10 in execute_code. 2023-04-03 11:34:39 +01:00
Toran Bruce Richards
c9be6edf9e Improves parsing of AI JSON output. 
This makes gpt3.5 turbo fully possible! 🚀
 2023-04-03 11:34:17 +01:00
Toran Bruce Richards
2bdd6b7b83 Adds missing google search import. 2023-04-03 11:33:35 +01:00
Toran Bruce Richards
ba6df3674b Fixes overwrite memory command. 2023-04-03 11:33:18 +01:00
Toran Bruce Richards
6a0a1355f0 Fixes google search command. 2023-04-03 11:33:01 +01:00
Toran Bruce Richards
b1cc474b92 removes unused commands. 2023-04-03 11:32:40 +01:00
Toran Bruce Richards
59d52e9bc7 Sets actual token limits. 2023-04-03 11:30:39 +01:00
Toran Bruce Richards
099a5e1090 Handles incorrect AI formatting in a more forgiving way. 2023-04-03 11:30:06 +01:00
Toran Bruce Richards
41daf07219 Vastly improves context window management. 
Now uses tokens and biggest context possible.
 2023-04-03 11:28:06 +01:00
Toran Bruce Richards
04710ae57b Implements counting of exact tokens 🚀 2023-04-03 11:26:09 +01:00
Toran Bruce Richards
5a669d96eb Evaluated code should not be JSON. 2023-04-03 11:25:43 +01:00
Toran Bruce Richards
1e73ee2958 Adds todo comment. 2023-04-03 11:24:36 +01:00
Taylor Brown
7fd2ce2bc6 Clean up where last_run_settings go 2023-04-02 22:12:14 -05:00
Taylor Brown
4173b07bce Remove excessive debug text 2023-04-02 22:12:12 -05:00
Taylor Brown
afceca0625 Remove commented code from ai_config 2023-04-02 22:11:06 -05:00
Toran Bruce Richards
0c1636565f Fixes missing openai import 2023-04-03 03:57:51 +01:00
Toran Bruce Richards
aed8c5b469 Fixes openai key not set error. 2023-04-03 03:57:38 +01:00
Taylor Brown
30f8ed95e1 Remove unused imports 2023-04-02 21:52:50 -05:00
Taylor Brown
ae9448cb89 Consolidate calls to openai 
Starting to abstract away the calls to openai
 2023-04-02 21:51:07 -05:00
Taylor Brown
744c5fa25b Remove hardcoding of gpt-3.5-turbo in favor of config 2023-04-02 21:38:21 -05:00
Taylor Brown
80ccd10d0b Use gpt-4 by default for the main thought process 
Allow specifying the llm through dotenv
Move more things into config
 2023-04-02 21:35:28 -05:00
Taylor Brown
3e587bc7fb Make the json_parser more robust 
For some reason the bot keeps prefacing its JSON. This fixes it for now.
 2023-04-02 18:50:51 -05:00
Taylor Brown
28c18adff0 Merge master to fixes_gpt3 2023-04-02 18:30:28 -05:00
Toran Bruce Richards
64af11f4ab Merge pull request #32 from eltociear/patch-1 
fix typo in browse.py
 2023-04-03 10:52:40 +12:00
Taylor Brown
10833c86b7 Add pyyaml package to requirements 2023-04-02 17:37:30 -05:00
Taylor Brown
dca5b78bfe Add missing json_parser.py and add alternative requirements 
This adds fix_and_parse_json
Also, add requirements-alternative.txt to help install reqs in a different environment
 2023-04-02 17:34:11 -05:00
Taylor Brown
2b5a7cc485 Add missing properties to config.py 2023-04-02 17:17:46 -05:00
Taylor Brown
9ff7e5954b Make compatible with gpt-3.5 
I made the json parsing more forgivable. I improved the prompt, using things I learned from: Koobah/Auto-GPT
 2023-04-02 16:39:17 -05:00
Taylor Brown
f808710528 Attempt to improve JSON handling in GPT-3 2023-04-02 14:48:35 -05:00
Ikko Eltociear Ashimine
0d94711d76 fix typo in browse.py 
specifc -> specific
 2023-04-03 00:34:01 +09:00
James C. Palmer
5e746afef8 Use load_dotenv() in chat.py and speak.py. 2023-04-02 09:32:25 -04:00
Toran Bruce Richards
99a33023c4 Fixes incorrect handling of openai RateLimitExceeded errors. 2023-04-02 13:46:31 +01:00
James C. Palmer
ef656a0f77 Remove keys.py and replace with python-dotenv. 
- Removed `keys.py`.
- Added `.env.template`.
- Added `.env` to `.gitignore`.
- Updated various files that imported `keys` to use `os.getenv` instead.
- Updated `requirements.txt` dependencies.
- Updated README.md with instructions on setting up environment variables.

This change improves security, flexibility, and makes it easier to use Auto-GPT in notebooks. Environment variables are stored in `.env` and loaded via `load_dotenv()` in `scripts/main.py`.
 2023-04-02 06:39:55 -04:00
James C. Palmer
cffe083ff2 Move requirements.txt to root directory. 2023-04-02 05:44:45 -04:00