53 Commits

Author SHA1 Message Date
peterychang
ba063e78ce disable youtube test until it can be debugged (#380) 2024-08-20 11:08:56 -07:00
Jack Gerrits
853b00b0f0 Add message context to message handler (#367)
Co-authored-by: Eric Zhu <ekzhu@users.noreply.github.com>
2024-08-17 03:14:09 +00:00
gagb
698a8f374b Catch json decode error in file surfer (#352)
* Remove commented code

* Removed unused code

* Add try except to catch json decode error

---------

Co-authored-by: afourney <adam.fourney@gmail.com>
2024-08-08 23:10:31 -07:00
afourney
365d21b55e Fixed numerous issues with orchestrator, related to 342 (#344)
* Fixed numerous issues with orchestrator, related to 342'

* Refined prompts, and added max wall-clock time to orchestrators.

* Fixed formatting.
2024-08-08 11:15:57 -07:00
Enhao Zhang
91e55d8a16 check ledger formatting (#327)
* check ledger formatting

* format file

* format file

* rename agent metadata field 'name'-->'type'

---------

Co-authored-by: Enhao Zhang <t-enhaozhang@microsoft.com>
2024-08-07 17:01:07 -07:00
afourney
c7f5931dc8 Reverts TeamOne prompts to those used in Autogen experiments. (#331)
* Revised prompts to match autogen experiments.

* Handle sh code blocks

* Move executor prompt into coder.py

* Fixed formatting.
2024-08-07 13:52:45 -07:00
Jack Gerrits
39489ba395 Rename fields in agent metadata (#336)
* Rename fields in agent metadata

* team one fixes

* another fix
2024-08-07 13:08:13 -07:00
afourney
ecde17a255 Re-added multimodality to mdconvert. (#330)
* Re-added multimodality to mdconvert.

* Fixed formatting.
2024-08-07 07:14:47 -07:00
afourney
8b13d59b59 Initial work porting WebArena to async (#325)
* Initial work porting webarena to async

* Perhaps resolved some of the eval() issues.
2024-08-06 16:25:18 -07:00
afourney
5ff275fad9 Handle errors in the TeamOne agent's message queue handler. (#324) 2024-08-05 13:36:25 -07:00
Andreas Volkmann
efa87784cc Validate agent description, DRY metadata (#321)
* Validate desc, DRY

* Update python/src/agnext/core/_base_agent.py

---------

Co-authored-by: Eric Zhu <ekzhu@users.noreply.github.com>
2024-08-05 18:07:19 +00:00
peterychang
136af65b74 Code execute cancellation (#299)
* Hook cancelation token into code execution

* Add unit test for code cancellation

* actually save the merge
2024-07-29 23:46:04 -07:00
afourney
ec654253d2 Introduces a BaseWorker agent, allowing for a TeamOneBaseAgent (#289) 2024-07-29 13:09:31 -07:00
peterychang
070a97ceaa make code execution async (#219)
* make code execution async

* python 3.10 does not support asyncio.timeout()

* make code execution cancellable

* make code execution async

* python 3.10 does not support asyncio.timeout()

* make code execution cancellable

* make entire callstack for code_executor async

* Update python/src/agnext/components/code_executor/_impl/local_commandline_code_executor.py

Co-authored-by: Jack Gerrits <jackgerrits@users.noreply.github.com>

* fix variable description

* remove unnecessary code

* fix usage of execute_code_blocks

* fix usage of execute_code_blocks

---------

Co-authored-by: Jack Gerrits <jackgerrits@users.noreply.github.com>
Co-authored-by: Eric Zhu <ekzhu@users.noreply.github.com>
2024-07-26 22:37:34 +00:00
Jack Gerrits
6437374f63 Handle messages in a queue for team one base agent (#284)
* Handle messages in a queue for team one base agent

* fmt

* Enable RPC

* Hatch errors.

---------

Co-authored-by: Adam Fourney <adamfo@microsoft.com>
Co-authored-by: afourney <adam.fourney@gmail.com>
2024-07-26 16:48:39 -04:00
afourney
99b53557ca Allow Executor to search for code in multimodal messages. (#285) 2024-07-26 12:54:10 -07:00
peterychang
cf2bcd31ea Web surfer test (#248)
* Add web surfer test

* delete temp files

* formatting

* fix send_message calls

* fix mypy errors

* Add web surfer test

* delete temp files

* formatting

* fix send_message calls

* fix mypy errors

* fix CI checks

* CI code formatting

* Update hatch commands, add tests to CI

* add playwright to test env

* try fixing toml

* Update .github/workflows/checks.yml

Co-authored-by: Jack Gerrits <jackgerrits@users.noreply.github.com>

* try fixing toml

* try fixing toml

* Update python/pyproject.toml

Co-authored-by: Jack Gerrits <jackgerrits@users.noreply.github.com>

* try fixing toml

* try fixing toml

* modify correct pyproject.toml file

* add missing dependency

* Add browser_utils tests

* fix check errors

* run normalize code for test

* add missing dependency

---------

Co-authored-by: Jack Gerrits <jackgerrits@users.noreply.github.com>
Co-authored-by: afourney <adam.fourney@gmail.com>
2024-07-26 12:34:47 -07:00
afourney
61dec0a3ca Added token counting to TeamOne (#264)
* Added token counting to TeamOne
2024-07-26 09:45:41 -07:00
afourney
25b9a91328 Teamone webarena agents fully implemented. (#276)
* Initial onboarding of WebArena
* Migrate to try_get_underlying_agent_instance
* Fixed some race conditions and implemented Deactivate messages.
2024-07-26 09:42:12 -07:00
afourney
9fd7e301b3 TeamOne agents default to handling one message at a time, exclusively… (#273)
* TeamOne agents default to handling one message at a time, exclusively. The default can be overridden by passing a parameter to the BaseAgent constructor.
2024-07-25 21:13:08 -07:00
Victor Dibia
782503dff5 Update team one readme with information on how logging works. (#259)
Update team one readme with information on how logging works.

Related to #228
2024-07-24 13:18:28 -07:00
Jack Gerrits
a52d3bab53 Agent factory can be async (#247) 2024-07-23 14:49:38 -04:00
gagb
e67ab18344 Improve team-one doc based on behavior (#237)
* Clarify generalist behav; add fig

* Add instructions for setting bing api key

* Add orchestration txt

* Add img for orch

* fix txt

* Add example for GAIA

* Add example for webarena

* answer q1

* answer q3

* Add what it cannot do

* add section on safety and risk
2024-07-19 01:25:52 +00:00
afourney
211bfa01c3 TeamOne implementation of GAIA (#221)
Port of GAIA benchmark
2024-07-17 09:51:19 -07:00
gagb
e69dd92c4f Improve team-one readme (#225)
* Update readme

* Improve readme further

* Add results
2024-07-16 22:18:06 +00:00
gagb
9e814cbad8 Update readme (#224)
* Update readme

* Improve readme further
2024-07-16 21:43:29 +00:00
afourney
524a5a78f9 Initial work moving mdconvert, and enabling browser Q&A (#211)
* Initial work moving mdconvert, and enabling browser Q&A
2024-07-15 09:47:38 -07:00
gagb
386680fd3e Add initial base and ledger orchestrator, improve executor (#206)
* Add initial base orchestrator

* Add basic ledger based orchestrator

* Improve executor; Improve prompts and logging for orchestration
2024-07-15 09:42:28 -07:00
Victor Dibia
78b63438e9 extend base log handler in team one to also write events to a log file (#212)
Looks good to me.
2024-07-12 22:21:45 +00:00
gagb
4c87ed71eb Improve file surfer (#204)
* Move file surfer to use baseagent

* Remove unused code

* Fix hatch err

* Intermediate commit to move mdconvert

* Fix linting errors

* Implement file surfer from autogen

* Run hatch check

* Fix more linting errors

* Fix errors in requests_markdown_browser. Ignore errors in mdconvert.

---------

Co-authored-by: Adam Fourney <adamfo@microsoft.com>
2024-07-12 12:48:47 -07:00
afourney
04d9709aec Agnext websurfer (#205)
* Initial work on multimodal websurfer

* A little more progress.

* Getting function calling to work.

* Some basic progress with navigation.

* Added ability to print multimodal messages to console.

* Fixed hatch error

* Nicely print multimodal messages to console.

* Got OCR working.

* Fixed the click action.

* Solved some hatch errors.

* Fixed some formatting errors.

* Fixed more type errors.

* Yet more fixes to types.

* Fixed many type errors.

* Fixed all type errors. Some needed to be ignored. See todos.

* Fixed all? hatch errors?

* Fixed multiline aria-names in prompts.
2024-07-11 10:52:29 -07:00
afourney
5996b452eb TeamOne handle multimodal console (#200)
* Added ability to print multimodal messages to console.

* Fixed hatch error
2024-07-10 00:01:13 -07:00
gagb
4553bf9054 Move file surfer to use baseagent (#198)
* Move file surfer to use baseagent

* Remove unused code

* Fix hatch err
2024-07-09 21:40:07 +00:00
afourney
ea2fcc28c3 Removes a redundant time field. (#197) 2024-07-09 14:02:09 -07:00
afourney
699f024a6d Adds a standard logging / log-printing class to TeamOne (#194)
* Added initial code for TeamOne utils.

* Fixed hatch errors.

* Updated examples.

* Fixed more hatch errors.

* examples/example_coder.py

* Added standard logging for TeamOne

* Read time from log record.
2024-07-09 13:51:05 -07:00
afourney
f7357499be Teamone utils (#192)
* Added initial code for TeamOne utils.

* Fixed hatch errors.

* Updated examples.

* Fixed more hatch errors.

* Improve readme

---------

Co-authored-by: gagb <gagb@users.noreply.github.com>
2024-07-09 17:46:55 +00:00
afourney
beada0279c Initial proof-of-concept for TeamOne BaseAgent class (#181)
* Initial proof-of-concept for TeamOne BaseAgent class

* Fixed most hatch errors.

* Handle final Hatch error

---------

Co-authored-by: gagb <gagb@users.noreply.github.com>
2024-07-08 22:36:25 +00:00
gagb
7d12c70cff Add missing file surfer dependency (#191) 2024-07-08 22:21:54 +00:00
afourney
8eb8a4b14d Added user proxy. (#176)
* Added user proxy.

* Add dependency

---------

Co-authored-by: gagb <gagb@users.noreply.github.com>
2024-07-03 17:13:24 -07:00
gagb
f82f3852d3 Update to use new src (#177) 2024-07-03 22:47:34 +00:00
afourney
99ecb5ec7f Added 'request_halt' flag, and fixed an orchestration bug (#174)
* Added 'request_halt' flag, and fixed an orchestration bug

* Fixed formatting errors.

* Fixed a hatch error with casting.
2024-07-03 00:04:44 -07:00
gagb
9df928b73e Improve printing further (#172)
* Improve pretty printing

* Improve printing further
2024-07-02 18:09:45 -07:00
gagb
e0065cab42 Improve pretty printing (#171) 2024-07-03 00:50:08 +00:00
gagb
bcec0502d7 Numerous fixes for agbench (#170)
* Shift to new runtime API

* Add pretty printing

* Reformat

* Fix linting errors
2024-07-02 23:18:48 +00:00
Jack Gerrits
766635394a Update send_message to be a single async operation. Add start helper to runtime to manage this (#165) 2024-07-01 11:53:45 -04:00
gagb
aff179776c Add coder, file surfer, common message types in team-one (#142)
* Update message types

* Add new coder, refactor file surfer

* Run hatch check

* Run hatch check
2024-06-28 15:50:10 -07:00
Jack Gerrits
13b0d0deb4 Implement closure agent (#143) 2024-06-28 10:22:44 -04:00
gagb
8901b4d224 Add module for common messages (#133)
* Move agents to a subdir

* Add placeholder for example team

* Move messages to a common file

* Add rounrobin orchestrator + reflex agents

* Fix import errors

* Clean up commented code

* Fix formatting errors

* Fix a linting error

* Fix formatting errors
2024-06-27 16:39:16 -07:00
Jack Gerrits
a13c971b16 Change send/publish api to better support async and represent reality (#137)
* Make send and publish better represent reality

* fix team-one
2024-06-27 13:40:12 -04:00
gagb
1cd39691f1 Move agents to a subdir (#128) 2024-06-25 14:50:42 -07:00