fix: resolve 4 bugs found via code audit

[sjhddh]

Summary

Fixes four bugs discovered during a code audit of the face swapper and UI modules:

1. Missing logging import (face_swapper.py)

• logging.error() is called in pre_check() but logging was never imported
• Causes NameError when model directory creation fails

2. Model filename mismatch (face_swapper.py)

• pre_check() downloads inswapper_128.onnx (FP32)
• get_face_swapper() loads inswapper_128.onnx (FP32)
• But pre_start() checks for inswapper_128_fp16.onnx — a file that is never downloaded
• This causes pre_start() to always fail with "Model not found"

3. VideoCapture resource leak (ui.py)

• render_video_preview() only calls capture.release() on the failure path
• When a frame is successfully read, the function returns without releasing the capture
• Leaks file handles on every video preview render

4. get_one_face() called with wrong arity (face_swapper.py)

• Line 530 passes two arguments (processed_frame, detected_faces) but get_one_face() only accepts one
• Causes TypeError at runtime in the map_faces fallback path
• Fixed by selecting from the already-detected faces directly, which also avoids redundant face detection

Test plan

• Verify pre_check() handles directory creation errors without NameError
• Verify pre_start() finds the correct model file after pre_check() downloads it
• Verify video preview doesn't leak file handles (check with lsof during preview)
• Verify map_faces fallback path works without TypeError

🤖 Generated with Claude Code

Summary by Sourcery

Fix bugs in the face swapper model handling and face selection logic, and ensure video preview resources are correctly released.

Bug Fixes:

• Ensure face swapper startup checks for the same ONNX model file that is downloaded during pre-check.
• Fix the face swap fallback path by selecting from already-detected faces instead of calling a helper with an invalid signature.
• Release video capture resources on successful video preview rendering to avoid leaking file handles.

[sourcery-ai]

Reviewer's guide (collapsed on small PRs)

Reviewer's Guide

Fixes four audited bugs in the face swapper and UI modules: corrects the face swapper model path and a wrong-arity face selection call, and ensures proper logging import and video capture cleanup in the UI preview flow.

Sequence diagram for face swapper model pre_check and pre_start flow

sequenceDiagram
    participant App
    participant FaceSwapper as face_swapper_module
    participant FileSystem

    App->>FaceSwapper: pre_check()
    FaceSwapper->>FileSystem: ensure models_dir exists
    alt models_dir creation fails
        FileSystem-->>FaceSwapper: OSError
        FaceSwapper->>FaceSwapper: logging.error(error)
        FaceSwapper-->>App: False
    else models_dir ok
        FileSystem-->>FaceSwapper: ok
        FaceSwapper-->>App: True
    end

    App->>FaceSwapper: pre_start()
    FaceSwapper->>FaceSwapper: model_path = join(models_dir, inswapper_128.onnx)
    FaceSwapper->>FileSystem: exists(model_path)
    alt model file exists
        FileSystem-->>FaceSwapper: True
        FaceSwapper-->>App: True
    else model file missing
        FileSystem-->>FaceSwapper: False
        FaceSwapper->>FaceSwapper: update_status(Model not found: inswapper_128.onnx)
        FaceSwapper-->>App: False
    end

Loading

Sequence diagram for render_video_preview VideoCapture lifecycle

sequenceDiagram
    actor User
    participant UI as ui_module
    participant CV2 as cv2_VideoCapture

    User->>UI: render_video_preview(video_path, size)
    UI->>CV2: VideoCapture(video_path)
    UI->>CV2: read()
    alt frame read successful
        CV2-->>UI: ret=True, frame
        UI->>UI: gpu_cvt_color(frame)
        UI->>UI: Image.fromarray(...)
        alt size provided
            UI->>UI: ImageOps.fit(image, size, LANCZOS)
        end
        UI->>CV2: release()
        UI-->>User: CTkImage
    else frame read failed
        CV2-->>UI: ret=False, None
        UI->>CV2: release()
        UI->>UI: cv2.destroyAllWindows()
        UI-->>User: None or error image
    end

Loading

File-Level Changes

Change	Details	Files
Align face swapper pre_start() model path with the model actually downloaded in pre_check().	Change pre_start() to look for inswapper_128.onnx instead of the non-downloaded inswapper_128_fp16.onnx Keep failure messaging and existence check logic the same so behavior only changes in which file is required	modules/processors/frame/face_swapper.py
Fix incorrect use of get_one_face() in the map_faces fallback path by reusing already-detected faces directly.	Replace the call to get_one_face(processed_frame, detected_faces) with a direct selection of a target face from detected_faces using the smallest x-coordinate (left-most face) Guard face selection with a truthy detected_faces check and only append source/target pairs when both faces are present	modules/processors/frame/face_swapper.py
Ensure video preview releases VideoCapture on both success and failure paths to avoid resource leaks.	Call capture.release() immediately before returning the successfully created CTkImage in render_video_preview() Keep the existing release and cv2.destroyAllWindows() calls on the failure path so all paths clean up resources	modules/ui.py

Possibly linked issues

• #(not specified): PR fixes the inswapper filename mismatch in face_swapper.py that causes the model load failure described in issue.

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey - I've left some high level feedback:

• In render_video_preview, consider wrapping cv2.VideoCapture usage in a context manager or a try/finally block so the capture is reliably released even if an exception occurs between opening the stream and returning the image.
• The fallback face selection now uses min(detected_faces, key=lambda x: x.bbox[0]); if get_one_face encodes the canonical selection heuristic, it may be better to reuse or refactor that logic so face-selection behavior stays consistent in one place and does not rely on assumptions about bbox.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- In `render_video_preview`, consider wrapping `cv2.VideoCapture` usage in a context manager or a try/finally block so the capture is reliably released even if an exception occurs between opening the stream and returning the image.
- The fallback face selection now uses `min(detected_faces, key=lambda x: x.bbox[0])`; if `get_one_face` encodes the canonical selection heuristic, it may be better to reuse or refactor that logic so face-selection behavior stays consistent in one place and does not rely on assumptions about `bbox`.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}