core-conversion
Convert files, streams, paths, URIs, and HTTP responses to Markdown with the MarkItDown Python API and CLI.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- core-conversion
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Convert files, streams, paths, URIs, and HTTP responses to Markdown with the MarkItDown Python API and CLI.
- disable-model-invocation
- true
02 资源文件 5
- 参考文档references/api-reference.md6.3 KB
- 参考文档references/cli-reference.md3.3 KB
- 参考文档references/formats-and-dependencies.md3.7 KB
- 参考文档references/troubleshooting.md4.0 KB
- 脚本scripts/convert_smoke.py5.5 KB
03 SKILL.md 正文
Core Conversion
Use this sub-skill when a task needs offline MarkItDown conversion through the Python API or markitdown CLI: local files, Path objects, binary streams, file:, data:, http: or https: URIs, and requests.Response objects.
Start Here
- For Python entry points, result objects, stream hints, URI handling, and exception handling, read references/api-reference.md.
- For shell usage, stdin conversion, output files, stream hints, plugin listing, and data URI preservation, read references/cli-reference.md.
- For built-in converter coverage and optional dependency extras, read references/formats-and-dependencies.md.
- For unsupported formats, missing extras, failing converters, I/O boundaries, stream buffering, audio tool warnings, and large embedded data, read references/troubleshooting.md.
- To smoke-test the installed package safely from any directory, run
python scripts/convert_smoke.py --check-importorpython scripts/convert_smoke.pyfrom this sub-skill directory.
Routing Boundaries
- Use this sub-skill for
MarkItDown,StreamInfo,DocumentConverterResult,DocumentConverter,convert,convert_local,convert_stream,convert_uri,convert_response, themarkitdownCLI, built-in converters, local fixtures, and conversion errors. - Route Azure Document Intelligence or Azure Content Understanding conversion to
../cloud-integrations/SKILL.md. - Route OCR plugin usage to
../ocr-plugin/SKILL.md. - Route custom converter or plugin authoring to
../plugin-development/SKILL.md. - Route MCP serving to
../mcp-server/SKILL.md. - Do not make real LLM, cloud, or external API calls from this sub-skill; document required kwargs only.
Safe Conversion Pattern
- Choose the narrowest API:
convert_localfor trusted local paths,convert_streamfor caller-owned bytes,convert_urionly when URI fetching or URI decoding is intended, andconvert_responsewhen an HTTP client already fetched the response. - Add
StreamInfo(extension=..., mimetype=..., charset=...)when bytes have no reliable filename, content type, or charset. - Capture
result.markdown; useresult.titlewhen a converter provides a title, andresult.text_contentonly for compatibility with older code. - Preserve embedded data URIs only when explicitly needed with
keep_data_uris=Trueor CLI--keep-data-uris. - Catch
MissingDependencyException,UnsupportedFormatException, andFileConversionExceptionseparately so users know whether to install an extra, choose another tool, or inspect converter-specific failures.
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).