406 Commits

Author SHA1 Message Date
Xingyao Wang
84a578ad20
[test] remove integration tests from CI & move them into evaluation (#4447) 2024-10-17 05:38:23 +08:00
mamoodi
6f2e678028
Fix eval output path in case of @ char (#4416) 2024-10-15 22:45:08 +00:00
Abhijeetsingh Meena
173018eb58
fix: Resolves HumanEval Inference by replacing task_id with instance_id (#4364)
Co-authored-by: Harshit Surana <surana.h@gmail.com>
2024-10-15 15:18:38 +00:00
Xingyao Wang
50c13aad98
[Eval] Improve SWE-Bench Eval harness: multi-run support & entry script simplification (#4396) 2024-10-15 21:34:52 +08:00
Xingyao Wang
25f9413965
[Eval] Fix eval stuck when result is too large for pbar (#4361) 2024-10-14 22:08:34 +08:00
Xingyao Wang
4dfc7a7ef0
[Eval] Add a more lightweight / easier-to-use SWE-Bench output visualizer (#4360)
Co-authored-by: Engel Nyst <enyst@users.noreply.github.com>
2024-10-14 02:09:01 +00:00
Xingyao Wang
b23c7aab5a
[eval] stop set sid in eval (#4311) 2024-10-10 11:47:27 +08:00
Robert Brennan
45fb4fb9bc
allow reconnecting to a runtime (#4223) 2024-10-09 16:37:52 +00:00
Engel Nyst
e6847e9e61
Move agenthub within openhands (#4130) 2024-10-08 00:34:18 +00:00
Alejandro Cuadron Lafuente
a3571ec510
[Fix] Error when trying to pull all docker evaluation containers (#4244) 2024-10-08 05:03:36 +08:00
Aditya Bharat Soni
0809d26f4d
fix: Allow evaluation benchmarks to pass image urls in run_controller() instead of simply passing strings (#4100)
Co-authored-by: Xingyao Wang <xingyao@all-hands.dev>
2024-10-07 15:37:08 -04:00
Xingyao Wang
01ae54a69d
fix swebench repo/version being string (#4241) 2024-10-07 22:01:42 +08:00
Xingyao Wang
245334e89d
[eval] improve update output script for swe-bench (#4180) 2024-10-04 15:10:03 +00:00
Xingyao Wang
80a631361b
eval: update aiderbench readme (#4209) 2024-10-04 09:26:12 -04:00
Xingyao Wang
9cc9b19958
eval: improve swebench infer error handling and retry (#4205) 2024-10-04 07:09:56 -05:00
Xingyao Wang
0c2a35b256
[eval] update aider bench scripts (#4203) 2024-10-04 02:23:06 +00:00
tofarr
152f99c64f
Chore Bump python version (#3545) 2024-10-03 13:40:55 -04:00
Xingyao Wang
53a015f718
fix: make llm_completions optional to fix eval_infer.py (#4148) 2024-10-02 03:55:03 +08:00
mamoodi
0144caaf1f
Update eval doc for remote runtime (#4145) 2024-10-01 13:14:36 -04:00
Xingyao Wang
1109637efb
Update instruction for new version of eval runtime-api (#4128) 2024-09-30 23:48:38 +00:00
Xingyao Wang
8d6eda3623
fix eval_infer.sh to correctly copy SWE-Bench logs (#4111) 2024-09-29 18:39:18 -05:00
tobitege
c3bbe604eb
(fix) Fix logging in shared eval file to prevent key disclosure (#4108) 2024-09-28 19:33:16 +00:00
Xingyao Wang
81b3cd71b3
[eval] log evaluating warnings directly to console (#4026) 2024-09-26 03:42:32 +08:00
Xingyao Wang
1b1d8f0b02
[eval] Use imap_unorderd for parallizing evaluation (#4040) 2024-09-24 20:47:27 +00:00
Xingyao Wang
a66e738957
[eval] use mp Pool instead ProcessPoolExecutor (#4025) 2024-09-24 23:59:06 +08:00
Ikko Eltociear Ashimine
c84495830e
[eval] update swe_bench/README.md (#3990) 2024-09-23 11:03:09 +02:00
Xingyao Wang
714e46f29a
[eval] save eventstream & llm completions for SWE-Bench run_infer (#3923) 2024-09-22 04:39:13 +00:00
Xingyao Wang
b13ed017d8
[eval] add git patch post-processing for SWE-Bench eval_infer (#3980) 2024-09-20 15:33:53 +00:00
Engel Nyst
8fdfece059
Refactor messages serialization (#3832)
Co-authored-by: Robert Brennan <accounts@rbren.io>
2024-09-18 23:48:58 +02:00
tofarr
ad0b549d8b
Feat Tightening up Timeouts and interrupt conditions. (#3926) 2024-09-18 20:50:42 +00:00
Xingyao Wang
5d7f2fd4ae
[eval] Allow evaluation of SWE-Bench patches on RemoteRuntime (#3927)
Co-authored-by: Boxuan Li <liboxuan@connect.hku.hk>
Co-authored-by: Graham Neubig <neubig@gmail.com>
2024-09-18 16:07:34 -04:00
Engel Nyst
ef09f0fe37
Small fix in readme (#3912) 2024-09-17 14:33:25 +00:00
Xingyao Wang
f996b31d64
[eval] Fix multi-processing bug (again^3) & allow set EXP_NAME for each run_infer (#3907)
Co-authored-by: Boxuan Li <liboxuan@connect.hku.hk>
2024-09-17 14:07:58 +00:00
tobitege
52c5abccbf
(enh) Dockerfile.j2: improve env vars for bash and activate in .bashrc (#3871) 2024-09-17 08:49:04 +02:00
Graham Neubig
243cb492aa
Run pre-commit on all files (#3884) 2024-09-16 11:07:08 -04:00
Xingyao Wang
2b3925278d
[eval] refactor process instance logic into update_progress (#3875) 2024-09-15 18:47:15 -04:00
Engel Nyst
379f2b6f23
Fix queue length on Macs (#3867) 2024-09-14 01:11:29 +00:00
Xingyao Wang
3a1b8c093b
[eval] yet another eval fixes on multi-processing (#3854)
Co-authored-by: Graham Neubig <neubig@gmail.com>
2024-09-13 15:51:22 +00:00
Xingyao Wang
78c5f58adc
refactor & improve retry for the reliability of RemoteRuntime & evaluation (#3846) 2024-09-13 07:37:07 -04:00
Xingyao Wang
797f02ff6f
rename huggingface evaluation benchmark (#3845) 2024-09-12 18:50:26 +00:00
Xingyao Wang
47d9621742
[eval] SWE-Bench eval usability fixes (#3836)
* [eval] increase timeout for swebench eval init/complete

* allow CmdRunAction to optionally block when .timeout is setted

* fix unit test for serialization

* fix unit tests for security analyzer

* fix integration tests

* add more timeout

* only check P2P when instances are non-empty;
convert P2P and F2P columns to string instead of list

---------

Co-authored-by: Graham Neubig <neubig@gmail.com>
2024-09-12 16:33:51 +00:00
Xingyao Wang
2fe2f4c530
[eval] increase timeout for SWEBench eval init/complete (#3829)
* [eval] increase timeout for swebench eval init/complete

* allow CmdRunAction to optionally block when .timeout is setted

* fix unit test for serialization

* fix unit tests for security analyzer

* fix integration tests

* add more timeout
2024-09-12 15:20:58 +00:00
Jiayi Pan
43c4a7fff4
Allow Generalized SWE-Bench format for evaluation (#3752)
* allow generalized swe-bench format

* Update run_infer.py

* fix linter

---------

Co-authored-by: Xingyao Wang <xingyao6@illinois.edu>
2024-09-06 13:05:00 +00:00
Xingyao Wang
688068a44e
Fix issues for running RemoteRuntime in parallel on SWE-Bench (#3716)
* feat: add SWE-bench fullset support

* fix instance image list

* update eval script and documentation

* increase timeout for remote runtime

* add push script

* handle the case when ret push is an generator

* update pbar

* set SWE-Bench default to run SWE-Bench lite

* add script to cleanup remote runtime

* fix the cases when tag is too long

* update README

* update readme for cleanup

* rename od to oh

* Update evaluation/swe_bench/README.md

Co-authored-by: Graham Neubig <neubig@gmail.com>

* Update evaluation/swe_bench/README.md

Co-authored-by: Graham Neubig <neubig@gmail.com>

* Update evaluation/swe_bench/scripts/cleanup_remote_runtime.sh

Co-authored-by: Graham Neubig <neubig@gmail.com>

* Update evaluation/swe_bench/scripts/cleanup_remote_runtime.sh

Co-authored-by: Graham Neubig <neubig@gmail.com>

* Update evaluation/swe_bench/scripts/cleanup_remote_runtime.sh

Co-authored-by: Graham Neubig <neubig@gmail.com>

* gets API key and Runtime from env var

---------

Co-authored-by: Graham Neubig <neubig@gmail.com>
2024-09-05 10:34:31 +08:00
Xingyao Wang
d8a87d7ccb
[Eval] Make SWE-Bench run_infer.sh to default to run SWE-Bench Lite (#3704)
* feat: add SWE-bench fullset support

* fix instance image list

* update eval script and documentation

* increase timeout for remote runtime

* add push script

* handle the case when ret push is an generator

* update pbar

* set SWE-Bench default to run SWE-Bench lite
2024-09-04 00:58:14 +08:00
Xingyao Wang
d283420ac2
feat: add SWE-bench fullset support (#3477)
* feat: add SWE-bench fullset support

* fix instance image list

* update eval script and documentation

* add push script

* handle the case when ret push is an generator

* update pbar
2024-09-02 20:28:52 -04:00
mamoodi
6fcc4ca052
fix eval README link (#3692) 2024-09-02 09:29:42 +08:00
tobitege
dbb671a8a5
logname fix; improve test calling instruction (#3666) 2024-08-30 17:15:31 +02:00
Xingyao Wang
090c911a50
(refactor) Make Runtime class synchronous (#3661)
* change runtime to be synchronous

* fix test runtime with the new interface

* fix arg

* fix eval

* fix missing config attribute

* fix plugins

* fix on_event by revert it back to async

* update upload_file endpoint

* fix argument to upload file

* remove unncessary async for eval;
fix evaluation run in parallel

* use asyncio to run controller for eval

* revert file upload

* truncate eval test result output
2024-08-30 01:37:03 +00:00
Xingyao Wang
8b1f207d39
feat: support remote runtime (#3406)
* feat: refactor building logic into runtime builder

* return image name

* fix testcases

* use runtime builder for eventstream runtime

* have runtime builder return str

* add api_key to sandbox config

* draft remote runtime

* remove extra if clause

* initialize runtime based on box class

* add build logic

* use base64 for file upload

* get runtime image prefix from API

* replace ___ with _s_ to make it a valid image name

* use /build to start build and /build_status to check the build progress

* update logging

* fix exit code

* always use port

* add remote runtime

* rename runtime

* fix tests import

* make dir first if work_dir does not exists;

* update debug print to remote runtime

* fix exit close_sync

* update logging

* add retry for stop

* use all box class for test keep prompt

* fix test browsing

* add retry stop

* merge init commands to save startup time

* fix await

* remove sandbox url

* support execute through specific runtime url

* fix file ops

* simplify close

* factor out runtime retry code

* fix exception handling

* fix content type error (e.g., bad gateway when runtime is not ready)

* add retry for wait until alive;
add retry for check image exists

* Revert "add retry for wait until alive;"

This reverts commit dd013cd2681a159cd07747497d8c95e145d01c32.

* retry when wait until alive

* clean up msg

* directly save sdist to temp dir for _put_source_code_to_dir

* support running testcases in parallel

* tweak logging;
try to close session

* try to close session even on exception

* update poetry lock

* support remote to run integration tests

* add warning for workspace base on remote runtime

* set default runtime api

* remove server runtime

* update poetry lock

* support running swe-bench (n=1) eval on remoteruntime

* add a timeout of 30 min

* add todo for docker namespace

* update poetry loc
2024-08-29 15:53:37 +00:00