Commit graph

14 commits

Author SHA1 Message Date
Daniel Han
797ddd201e
Fix Studio silently exiting on Windows without error output (#4527)
* Fix Studio silently exiting on Windows without error output

On Windows, `unsloth studio` launches a child process via
subprocess.Popen to run the server in the studio venv. If the child
crashes (e.g. due to a missing package), the parent just calls
typer.Exit(rc) with no message -- the user sees "Launching Unsloth
Studio... Please wait..." and then the prompt returns with zero
feedback.

Root cause: `data_designer_unstructured_seed` is imported at the top
level in seed.py. If this package is not installed in the studio venv,
the entire import chain (seed.py -> routes/__init__.py -> main.py ->
run_server()) crashes with ModuleNotFoundError. Since run.py has no
try/except around run_server() and studio.py does not report nonzero
exit codes, the failure is completely silent.

Changes:
- run.py: wrap run_server() in try/except, print clear error with
  traceback to stderr. Also reconfigure stderr encoding on Windows so
  tracebacks with non-ASCII paths do not cause secondary failures.
- studio.py: print an error message when the child process exits with
  a nonzero code on Windows, so the user knows something went wrong.
- seed.py: make data_designer_unstructured_seed import optional with
  a try/except fallback. The server starts normally and only returns
  HTTP 500 if the unstructured seed endpoints are actually called.

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Skip Anaconda/Miniconda Python when creating Studio venv on Windows

Conda-bundled CPython ships modified DLL search paths that prevent
torch from loading c10.dll on Windows. The Studio server fails
silently at startup because the venv was created with conda's Python.

Standalone CPython (python.org, winget, uv) does not have this issue.

Both install.ps1 and setup.ps1 now skip any Python binary whose path
contains conda, miniconda, anaconda, miniforge, or mambaforge when
selecting the interpreter for the studio venv. If only conda Python
is available, the scripts print an error with instructions to install
standalone CPython.

* Fix multi-file preview crash and improve setup.ps1 Python discovery

Addresses review findings [10/10] and [8/10]:

1. seed.py: _read_preview_rows_from_multi_files() had a hard import
   of build_multi_file_preview_rows inside the function body, bypassing
   the optional-plugin guard. Moved it into the top-level try/except
   block and added a None guard matching the other functions.

2. setup.ps1: Python discovery now probes py.exe (Python Launcher)
   first, uses Get-Command -All to look past conda entries that shadow
   standalone CPython further down PATH, skips WindowsApps stubs, and
   resolves the actual executable path so venv creation does not
   re-resolve back to a conda interpreter.

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Check sys.base_prefix to catch venvs created from conda Python

A venv created from conda Python (e.g. C:\Users\danie\.venv) has a
path that does not contain "conda", but sys.base_prefix still points
to the conda install (e.g. C:\Users\danie\miniconda3). The previous
path-only check missed this case entirely.

Both install.ps1 and setup.ps1 now use a Test-IsConda helper that
checks both the executable path AND sys.base_prefix against the
conda/miniconda/anaconda/miniforge/mambaforge pattern. This catches:
- Direct conda Python executables
- Venvs created from conda Python (base_prefix reveals the origin)

* Fix install.ps1 passing version string to uv venv instead of resolved path

Find-CompatiblePython returned a bare version string (e.g. "3.13")
which was passed to `uv venv --python 3.13`. uv performs its own
interpreter discovery and can resolve that version string back to a
conda Python, defeating the entire conda-skip logic.

Now Find-CompatiblePython returns a hashtable with both .Version (for
display) and .Path (the resolved absolute executable path). The venv
is created with `uv venv --python <absolute-path>`, ensuring uv uses
the exact interpreter we validated.

* Quote resolved Python path in uv venv call for paths with spaces

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2026-03-22 08:23:03 -07:00
Wasim Yousef Said
dd283b0605
feat(studio): multi-file unstructured seed upload with better backend extraction (#4468)
* fix(recipe-studio): prevent fitView from zooming to wrong location on recipe load

* feat: add pymupdf/python-docx deps and unstructured uploads storage root

* feat: add POST /seed/upload-unstructured-file endpoint

* feat: add multi-file chunking with source_file column

* feat: update frontend types and API layer for multi-file upload

* feat: round-robin preview rows across source files

Ensures every uploaded file is represented in the preview table
by cycling through sources instead of just taking the first N rows.

* fix: disable OCR, fix auto-load timing, fix persistence on reload

- Disable pymupdf4llm OCR with write_images=False, show_progress=False
- Replace onAllUploaded callback with useEffect that detects uploading→done
  transition (avoids stale closure reading empty file IDs)
- Fix importer to preserve file IDs from saved recipes instead of clearing
  (clearing only happens at share time via sanitizeSeedForShare)

* fix: harden unstructured upload with input validation and state fixes

Validate block_id/file_id with alphanumeric regex to prevent path
traversal, use exact stem match for file deletion, add error handling
for metadata writes and empty files, fix React stale closures and
object mutations in upload loop, and correct validation logic for
unstructured seed resolved_paths.

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix: address PR review - legacy path import, share sanitizer, sync effect

Promote legacy source.path into resolved_paths for old unstructured
recipes, clear source.paths in share sanitizer to prevent leaking local
filesystem paths, and gate file sync effect to dialog open transition
so users can actually delete all uploaded files.

* fix: CSV column fix (BOM + whitespace + unnamed index re-save) for #4470

* fix: harden unstructured upload flow and polish dialog UX

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2026-03-20 13:22:42 -07:00
Daniel Han
1f99dee027
fix(seed): disable remote code execution in seed inspect dataset loads (#4275)
* fix(seed): disable remote code execution for seed inspect loads

* fix(test): use __file__-relative path in seed test

The test used a CWD-relative path (`studio/backend/routes/...`) which
only resolved when pytest was invoked from the repo root. Use
`Path(__file__).resolve()` so the test passes regardless of CWD.

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: Test <test@test.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2026-03-13 19:37:43 +04:00
Roland Tannous
47654cb91c Final cleanup 2026-03-12 18:28:04 +00:00
Roland Tannous
a2baf80511 Update license headers 2026-03-12 17:23:10 +00:00
Shine1i
904e440513 feat(studio): studio storage roots path utilities 2026-03-11 20:19:52 +00:00
Roland Tannous
daa50d0756 Revert "Merge pull request #347 from unslothai/feature/studio-storage-roots"
This reverts commit 6b43e33ff1, reversing
changes made to 9edadaf21f.
2026-03-10 01:52:47 +00:00
Shine1i
5301514775 feat(studio): studio storage roots path utilities 2026-03-09 23:48:31 +00:00
Roland Tannous
d882678fe4 Add AGPL-3.0 SPDX headers to all source files 2026-03-09 20:17:45 +00:00
Shine1i
c67f4ba29f feat(recipe-studio): improve UI responsiveness and fix JSON preview handling 2026-03-09 16:04:46 +01:00
Shine1i
c88cce8185 refactor(seed): package unstructured seed reader as local Data Designer plugin 2026-03-03 11:22:04 +01:00
Shine1i
bdc825298d feat(seed): backend unstructured seed reader + server-side chunking, remove client chunk splitter 2026-03-03 11:11:26 +01:00
Shine1i
c3c65cded8 feat(recipe-studio): add image preview support for dataset and LLM configurations p1 2026-03-01 10:57:51 +01:00
Shine1i
e4b64f3cd5 refactor(data-recipe): split recipe backend routes for readability (seed/validate/jobs) 2026-02-26 14:05:32 +01:00
Renamed from studio/backend/routes/data_recipe.py (Browse further)