Documentation
These documents describe what raw2md does with a document and where you see the result of each stage.
Start with the reference. It describes the commands, the flags, the result files, the header, the service files, the exit codes, and the environment.
A document goes through the stages in the order of this table:
| Stage | When it runs | Document |
|---|---|---|
| Conversion | always, except for an md input |
reference.md |
| Cleaning | always, on every input | cleaning.md |
| Inspection | with --llm-inspection, on pdf and djvu only |
inspection.md |
| Post-processing | with --llm-post, on every input |
post.md |
| Quality evaluation | after cleaning, and again after each LLM stage | quality.md |
Each guide to a stage shows what the stage repairs or checks, with examples. It also shows what the stage does not change, and where you see its result.
Documents
- Reference
This reference describes the commands and flags of raw2md, the result files, the header, the service files, the exit codes, and the environment requirements.
- Cleaning
Cleaning is the first stage after conversion, and it always runs.
- Inspection
Inspection is an optional LLM stage after cleaning.
- Post-processing
Post-processing is an optional LLM stage, the last one before quality evaluation.
- Quality evaluation
Quality evaluation is a stage after cleaning, and it always runs.