What this category of data looks like
Data for this category primarily originates from the national petition information system, local petition systems, the 12345 citizen hotline platform, grassroots governance platforms, and historical petition archives, with daily updates. Commonly processed documents include petition registration forms, acceptance notices, transfer and assignment letters, handling opinions, review and re-examination applications, resolution plans, and similar items. Most are a mix of standardized official documents and structured forms, some include handwritten annotations or scanned archived items. Fields cover request content, handling entities, time limit nodes, review and re-examination records, and other relevant details, with no unified fixed format template.
Constraints on Document Parsing and Chunking
The diverse data sources include a mix of standardized files exported from systems and scanned documents, so the parsing process must support both text extraction and OCR adaptation. Daily updated batch processing requirements demand stable parsing workflows that can execute in bulk. Documents contain both structured form fields and free-form text content; chunking must retain the integrity of associated fields such as handling entities, requests, and time limits, to avoid breaking core handling logic. Historical petition archives have outdated files with inconsistent formats, so compatibility with multiple layout styles is required. Some documents contain sensitive handling information, so parsing must prevent leakage of sensitive fields.
Configuration Settings
These configurations apply to FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Adapts to the mixed length of standardized paragraphs and free-form text in petition documents, avoiding splitting core handling logic |
chunkSplitter | Split by paragraph + title | Matches the title hierarchy and paragraph structure of petition documents, retaining contextual connections of handling workflows |
PARSE_FILE_TIMEOUT_SECONDS | 300–600 seconds | Adapts to the time consumption requirements of batch scanned document parsing, avoiding timeout interruptions |
UPLOAD_FILE_MAX_SIZE | 100 MB | Compatible with the upload size of single petition archive packages, meeting batch archiving needs |
customPdfParse | Enable OCR mode | Adapts to parsing of paper archived documents such as scanned petition notices and handling opinions |
chunkSettingMode | Manual configuration | Flexibly adjust chunking rules for different types of petition documents, adapting to scenarios with mixed structured and unstructured content |
Three Common Mistakes
- Phenomenon: After deploying a Docker container, the GPU cannot be detected, and a memory overflow error is reported. Cause: GPU drivers and CUDA environment are not properly mounted, or a single task’s memory usage during batch parsing exceeds the hardware limit.
- Phenomenon: After importing Word-format petition documents, image links are lost when converting to Markdown format. Cause: Custom parsing rules to retain image domain names are not configured, and the chunking stage does not associate images with the original text context.
- Phenomenon: Timeout errors occur when parsing multi-page scanned petition opinions, and the task status shows failed. Cause: The
PARSE_FILE_TIMEOUT_SECONDSconfiguration value is too small to complete the full parsing process.
How to Verify Correct Configuration
- Upload a single petition registration form document, check the matching degree between the number of chunks in the parsing result and the number of original paragraphs, and adjust
chunkSizeto achieve the expected chunk length. - Batch upload 3 to 5 different types of petition documents, confirm that there are no timeout errors in parsing tasks, and verify whether the
PARSE_FILE_TIMEOUT_SECONDSconfiguration adapts to the current hardware environment. - Import a petition opinion document containing scanned content, check the integrity of the text after OCR parsing, and confirm that
customPdfParseis enabled. - Test uploading an archive package under 100 MB, confirm that the upload and parsing processes complete normally, and verify that
UPLOAD_FILE_MAX_SIZEmeets actual upload requirements.
The provided values are common starting points and should be evaluated against the user’s own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-28.