4fce28706f
SWE-agent Multimodal ( #1223 )
...
* Add ImageParsingHistoryProcessor
* Add SWEBenchMultimodal problem statement parsing
* cache processed swe-bench multimodal inputs
* Add testing for multimodal problem statement
* Update environment variable for PROBLEM_STATEMENT
* Update timeout params to multimodal config
* Add problem statement with images to Inspector view
* Update configs and add default_mm_no_images
* Add templates.disable_image_processing option
* Add Observation Images to Inspector
* Update configs for gemini
* Update docs for multimodal updates
* Fix: Protocols aren't inherited
* Chore: pathlib instead of os.path and make pre-commit happy
* add web_browser bundle
---------
Co-authored-by: carlosejimenez <cjsaltlake@gmail.com >
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2025-07-09 03:19:59 -04:00
ae8ded6f9e
Critical fix: WindowedFile last line display bug ( #1187 )
2025-06-02 14:36:34 -04:00
a42ec4fe97
Fix: Ignore missing treesitter for very old python ( #1180 )
...
Closes #1179
2025-05-31 14:10:57 -04:00
1effb7deca
Breaking: Remove review_on_submit (replaced by review_on_submit_m) ( #1148 )
2025-05-22 12:02:36 -04:00
1998ce1168
Breaking change: Rename defaults tool bundle ( #1147 )
...
* Breaking change: Rename defaults tool bundle
* Breaking change: Renaming of tool bundles
2025-05-22 11:54:45 -04:00
ce56deacda
Change(tools): Don't append \n to new file ( #1114 )
...
Co-authored-by: Boris Shigida <borshigida@gmail.com >
2025-05-10 12:50:11 -04:00
72a3680a71
Rename: ml_setup -> multilingual_setup
2025-05-01 10:19:30 +02:00
330cc464ce
Fix: Use multilingual install hack without adding as tool
2025-05-01 10:19:30 +02:00
aae4a22f4f
Support multilingual evaluation
2025-05-01 10:19:30 +02:00
fbf258ce48
Fix: Ensure state exists ( #1065 )
...
Previously, if other toolbundles were specified, the state might not
have existed (we try and read it before we even execute the first
action)
In reference to #1048
2025-04-08 12:06:11 -04:00
33adb47983
Fix: Ensure tools PATH takes precedence ( #1058 )
...
Closes #1057
2025-04-04 19:41:14 -04:00
163c03e319
Fix: Catch exit forfeit ( #1024 )
...
Closes #1011
2025-03-13 14:24:27 -04:00
1c445312ee
chore: update pre-commit hooks ( #1004 )
...
* chore: update pre-commit hooks
updates:
- [github.com/crate-ci/typos: v1.29.9 → v1.30.0](https://github.com/crate-ci/typos/compare/v1.29.9...v1.30.0 )
- [github.com/astral-sh/ruff-pre-commit: v0.9.7 → v0.9.9](https://github.com/astral-sh/ruff-pre-commit/compare/v0.9.7...v0.9.9 )
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* chore: Typo fix
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Kilian Lieret <kilian.lieret@posteo.de >
2025-03-07 17:38:57 -05:00
946a9fac2f
Tools: Delete diff state if diff state fails
2025-02-12 20:17:24 -08:00
616d79c47d
Feat(tools): Generalizatoin of review_submit
2025-02-12 20:17:24 -08:00
2100ba94fa
Fix(tools): Work around further unicode issues
...
django__django-13660/django__django-13660.debug.log
717:UnicodeEncodeError: 'utf-8' codec can't encode characters in
position 341-343: surrogates not allowed
django__django-10914/django__django-10914.debug.log
1638:UnicodeEncodeError: 'utf-8' codec can't encode characters in
position 193-195: surrogates not allowed
2025-02-12 20:17:24 -08:00
f17e14a41e
Fix(tools): UnicodeDecode errors in state/submit
2025-02-12 20:17:24 -08:00
fce933b077
Tools: Try enforcing utf-8 everywhere
2025-02-12 20:17:24 -08:00
20bad880b0
Fix(tools): Correct submit with review on submit
2025-02-12 20:17:24 -08:00
3296f8d95e
Feat(tools): Diff as state
2025-02-12 20:17:24 -08:00
4c99935ed1
Enh(tools/agent): State is written to file instead of extracted from
...
stdout
2025-02-12 20:17:24 -08:00
7d7e38dc3c
Enh(tools/agent): Read patch from file instead of stdout
2025-02-12 20:17:24 -08:00
7fef6cfc76
Enh(tools): Better warning if dir doesn't exist
2025-02-12 20:17:24 -08:00
1690f9b8ae
Fix(tools): Default to utf-8 encoding when writing
2025-02-12 20:17:24 -08:00
802af1753a
Fix(tools): Disable debug msgs with linting
2025-02-12 20:17:24 -08:00
1330d43559
Feat(tools): Bring back linter as warning
2025-02-12 20:17:24 -08:00
2996d8f84e
Fix(tools): Handling of @ for window expansion
2025-02-12 20:17:24 -08:00
43de78f30b
Fix(tools): Correct viewport expansion
2025-02-12 20:17:24 -08:00
1d1f367329
Fix(tools): Avoid breaking assertions
2025-02-12 20:17:24 -08:00
2190cf21fa
Fix(tools): Avoid shrinking the viewport
2025-02-12 20:17:24 -08:00
569e8e693a
Enh(tools): Admonish to use view lines after we show filemap
2025-02-12 20:17:24 -08:00
115195fcd3
Fix(tools): Do not include def line; handle decorators
2025-02-12 20:17:24 -08:00
29f751226d
Fix(tools): Solve encoding error
2025-02-12 20:17:24 -08:00
368ce556f1
Enh(tools): use utf-8 with errors=replace
2025-02-12 20:17:24 -08:00
07442337e4
Fix(tools): Attempt to work around unicode errors
2025-02-12 20:17:24 -08:00
69bdeaa0d9
Tools: Add filemap
2025-02-12 20:17:24 -08:00
f347484e02
Add WindowExpander to str_replace_editor
2025-02-12 20:17:24 -08:00
248c485223
Fix(tools): Don't attempt to edit if repl strs are the same
2025-02-12 20:17:24 -08:00
d5baa2da87
Fix(tools): Encoding issues
2025-02-12 20:17:24 -08:00
692aa8b9e7
Fix(tools): model.patch in patch breaks sb-cli eval
2025-02-12 20:17:24 -08:00
23f5a361b0
Enh(tools): 'create' creates folder if !exists
2025-02-12 20:17:24 -08:00
a95a2b1e1e
Fix: Retrieve problem statement for template
2025-02-12 20:17:24 -08:00
5f935edeb0
Feat: review on submit prototype
2025-02-12 20:17:24 -08:00
6f21d11ab9
Adjust prompts
2025-02-12 20:17:24 -08:00
04e12d5988
Fix: Description of scroll_up ( #857 )
...
the default action "scroll_up" has inconsistent description, the same as for "scroll_down"
2024-12-10 15:06:19 -05:00
25c5e1c616
Fix: Update tools after removing utils.sh
2024-12-09 12:25:14 -05:00
938cc4fb3b
Enh(tools): Fall back to env variables with registry.get
2024-12-09 10:11:02 -05:00
a2fcd6ee9d
CI: Fix unittests after introducing registry bundle
2024-12-07 16:25:26 -05:00
5c25314f47
Tools: Split out registry as separate bundle
2024-12-07 11:57:20 -05:00
b0536b29dd
Delete CTF tools that are currently incompatible
2024-12-07 11:57:20 -05:00
Kilian Lieret