layout-objects
Routes LayoutParser coordinate primitives, layout containers, shape operations, and layout-analysis helpers.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- layout-objects
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Routes LayoutParser coordinate primitives, layout containers, shape operations, and layout-analysis helpers.
- disable-model-invocation
- true
02 资源文件 1
- 参考文档references/guide.md3.9 KB
03 SKILL.md 正文
Layout Objects
Use this sub-skill for geometry-heavy LayoutParser tasks: coordinate conversion, padding, containment, serialization, sorting, and lightweight line or category grouping.
What belongs here
Interval,Rectangle,Quadrilateral,TextBlock, andLayout- relative/absolute coordinate transforms
pad,shift,scale,crop_imageunion,intersect,is_in,condition_on,relative_toto_dict,from_dict,to_dataframe,get_homogeneous_blocksgeneralized_connected_component_analysis_1dsimple_line_detectiongroup_textblocks_based_on_category
What does not belong here
- JSON/CSV/PDF loading and export: use
layout-io draw_box/draw_text: usevisualization- OCR response parsing: use
ocr lp://model catalogs or model inference: uselayout-models
Read these files
references/guide.mdfor the class, operation overview, and troubleshooting notes../layout-io/SKILL.mdwhen the task starts with files or PDFs../visualization/SKILL.mdwhen the task ends with rendering
Fast path
- Identify the shape type on each block.
- Use the most specific conversion first (
Interval↔Rectangle↔Quadrilateral). - Use
Layoutbatch methods for page-wide transforms. - Keep
TextBlockwrappers when the block also carries text or ids. - Only approximate quadrilateral shape operations with
strict=Falsewhen a bounding-box approximation is acceptable.
Common user requests
- "Filter the left column"
- "Convert boxes to rectangles"
- "Union the regions"
- "Check whether a token is inside a page block"
- "Sort blocks by reading order"
- "Group the text blocks by line or category"
- "Round-trip a layout to dict or dataframe"
Minimal smoke
Use the bundled root smoke script when you only need to confirm the geometry layer and serialization basics:
python ../../scripts/smoke_layoutparser_core.py
Failure clues
InvalidShapeError: the chosen union/intersection is geometrically invalid.NotSupportedShapeError: the operation would create a polygonal result that LayoutParser cannot represent directly.ValueErrorfromLayout(...): the input is not an iterable of blocks or is already aLayoutwrapper.ValueErrorfromTextBlock.to_interval(): the axis is missing.
Output discipline
When answering, give the exact class names and coordinate assumptions that the future agent should use. If a block should be re-encoded as another shape, state the destination shape explicitly so the user can follow the conversion.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).