cli-and-scripts
Use docTR's installed OCR CLI and safe bundled helper scripts for single-document and batch OCR.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- cli-and-scripts
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Use docTR's installed OCR CLI and safe bundled helper scripts for single-document and batch OCR.
- disable-model-invocation
- true
02 资源文件 6
- 参考文档references/cli-reference.md4.7 KB
- 参考文档references/script-adaptations.md6.5 KB
- 参考文档references/troubleshooting.md5.0 KB
- 脚本scripts/doctr_batch_ocr.py10.4 KB
- 脚本scripts/doctr_cli_env.py5.7 KB
- 脚本scripts/doctr_quick_ocr.py8.5 KB
03 SKILL.md 正文
CLI and Scripts
Use this sub-skill when the task is about running docTR from a shell instead of writing a custom Python pipeline.
Route here for
- Exact
doctr-cliarguments, defaults, JSON output behavior, and parser/runtime errors. - Single-file OCR with a safe helper that can run offline with
--no-pretrainedor load trained weights with--pretrained. - Batch OCR over a file or directory with explicit
txt,json, orxmloutput. - CLI/script troubleshooting: missing entry point, import failures, unsupported inputs, failed output writes, model-name errors, or model download/cache issues.
Runtime files
- Read references/cli-reference.md for the installed
doctr-clicommand, flags, defaults, output schema, and error behavior. - Read references/script-adaptations.md for the bundled helpers and how they differ from the repo-maintained examples.
- Read references/troubleshooting.md when commands fail.
- Run scripts/doctr_quick_ocr.py for one PDF or image.
- Run scripts/doctr_batch_ocr.py for a file or directory and
txt/json/xmloutputs. - Run scripts/doctr_cli_env.py for a safe CLI/import/backend diagnostic report.
Boundaries
- For equivalent Python API design, route to the core OCR/KIE sub-skill rather than expanding CLI snippets here.
- For document object schemas, rendering, hOCR/XML interpretation, Markdown/HTML export, and table/KIE result semantics, route to the document IO/export sub-skill.
- For training, evaluation, datasets, or benchmark scripts, route to the datasets/training/evaluation sub-skill.
Safety defaults
The bundled helpers default to --no-pretrained, which avoids intentional model-weight downloads and is useful for parser/import/offline smoke checks. Use --pretrained only when the runtime is allowed to load or download trained model weights and cache them through the installed backend.
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).