feat(filtering): add hough line detection operator - #389

Description

Adds a filtering_houghlines operator built on cv2.HoughLinesP, following the BaseOperator.compute shape of canny_edge.py.

• imagelab-backend/app/operators/filtering/hough_lines.py: converts the input to grayscale and runs Canny (fixed 50/150, aperture 3) on a copy, runs HoughLinesP on the edge map and draws the segments onto a copy of the original image. Reads rho, thetaDegrees, threshold, minLineLength, maxLineGap, color, thickness from self.params with defaults; thetaDegrees is converted to radians; color goes through hex_to_bgr. Out-of-range values raise ValueError with the same message style as contour_detection.py. Grayscale input is promoted to BGR and BGRA input keeps its alpha (lines drawn opaque), so the output channel count does not depend on whether lines were found. When HoughLinesP returns None the image is returned unchanged.
• imagelab-backend/app/operators/registry.py: registers filtering_houghlines next to filtering_cannyedge.
• imagelab-frontend/src/blocks/definitions/filtering.blocks.ts: new block with the seven fields, min/max limits matching the backend checks, and a tooltip noting the fixed Canny thresholds.
• imagelab-frontend/src/blocks/categories.ts: adds "Hough Line Detection" to the Filtering category.
• imagelab-frontend/src/data/operatorDocs.ts: docs entry for the operator and its parameters.

Only additions; no existing block types are renamed, so saved workspaces still load. The Canny thresholds are kept fixed because the issue scopes the block to the seven Hough/drawing parameters; exposing them is a follow-up if wanted.

Fixes #51

Type of Change

• Bug fix
• New feature
• Breaking change
• Documentation update

How Has This Been Tested?

Backend, imagelab-backend/tests/operators/filtering/test_hough_lines.py (31 tests): registry lookup; horizontal, vertical and 45 degree synthetic lines (3 px wide so Canny/Hough detection is deterministic) assert the drawn pixels lie within a few px of the known position and angle and span most of the line; segments are drawn on the original image, not the edge map; blank image returned unchanged; input not mutated; grayscale promoted to BGR with and without detections; BGRA keeps alpha; float input in [0,1]; thickness and colour honoured; a monkeypatched HoughLinesP test checks theta is passed in radians and the input is a binary edge map; 13 parameter validation cases plus invalid colour and unsupported shape; one round trip through POST /api/v1/pipeline/executions.

Frontend, imagelab-frontend/tests/blocks/houghLinesBlock.test.ts (5 tests): block defines the seven fields with the backend defaults, numeric fields clamp to the backend ranges, the block is listed in the Filtering category, docs cover every param, and extractPipeline on a real Blockly workspace emits the params by the names the backend reads.

Both files fail without the source change. Ran uv run ruff check ., uv run ruff format --check ., uv run pytest -q (822 passed; test_histogram_stats::test_calculate_channel_statistics_rgb fails on main as well, unrelated), npx eslint ., npx prettier --check, npx tsc -b, npx vitest run (212 passed).

Manual: ran backend and frontend locally, built Read image -> Hough Line Detection on a synthetic image with four lines and a rectangle, ran the pipeline and checked the processed output (screenshots below).

• Existing tests pass
• New tests added
• Manual testing

Screenshots (if applicable)

Pipeline and output after running:

Operator input and output:

Checklist

• My code follows the project's style guidelines
• I have performed a self-review
• I have added/updated documentation as needed
• My changes generate no new warnings
• Tests pass locally