What typical data for this category looks like
Typical data for this category comes from daily work outputs of disease control institutions, public health departments of medical institutions, and disease control divisions of health and health commissions. Daily documents include infectious disease report cards, epidemic analysis reports, case correction records, clustered epidemic reports, and epidemiological investigation reports. Data sources include the Chinese Center for Disease Control and Prevention Information System, hospital HIS systems, infectious disease reporting systems, and laboratory testing systems. Update frequency is daily. Documents fall into three categories: standardized forms, long-text analysis, and structured records. Core fields include legal infectious disease classification, case number, reporting time limit, correction notes, and similar items. There is no unified unstructured format.
What constraints do these characteristics impose on the document parsing and chunking link
Batch documents updated daily require the parsing process to be timely, to avoid delaying daily epidemic summary due to parsing delays. Scenarios with multiple document types require parsing tools to adapt to the different structures of standardized forms and long texts, to avoid losing field associations during chunking. The accuracy of core fields directly relates to the compliance of epidemic reports. The parsing process must retain the correspondence between fields and the main document body, and cannot tolerate omission of key information. Document formats vary across sources, so custom parsing rules must be supported to adapt to the standardized PDF formats used for disease control reporting, to avoid data loss caused by the default parsing logic failing to recognize specialized fields.
How to set up configurations
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Covers the core fields of a single case report or correction record, avoiding cutting off the association of case information during chunking |
chunkSplitter | Split by paragraph + heading | Matches the hierarchical structure of epidemiological investigation reports and epidemic analysis reports, retaining the integrity of report chapters |
customPdfParse | Enabled | Adapts to standardized PDF documents submitted for disease control reporting, retains the field correspondence of table cells, and solves the problem of abnormal recognition of standardized tables |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Addresses parsing time for daily updated batch documents, avoiding task interruption caused by parsing timeout of a single long epidemiological investigation report |
UPLOAD_FILE_MAX_SIZE | 100 MB | Accommodates a single complete clustered epidemic report or multi-page epidemiological investigation document |
chunkSettingMode | Manual configuration | Adjust chunking parameters for different document types (report cards, analysis reports) to adapt to scenario requirements |
Three common misconfigurations
- Phenomenon: After uploading a standardized infectious disease report card PDF, table fields in the parsing result are empty or misaligned. Cause: The
customPdfParseconfiguration is not enabled, and the default PDF parsing does not retain the cell field mapping of disease control standardized tables. - Phenomenon: When batch uploading documents updated on the same day, some parsing tasks show timeout failures. Cause: The
PARSE_FILE_TIMEOUT_SECONDSsetting value is too low, and does not adapt to the parsing time of a single long epidemiological investigation report. - Phenomenon: After uploading a multi-page epidemic analysis report, chunking results are scattered and cannot be associated with case correction records. Cause: The
chunkSizesetting is too small, cutting off cross-page case information association during chunking.
How to confirm the configuration is correct
- Upload a standard infectious disease report card PDF, check whether the parsing result retains core fields such as case number, report date, and infectious disease classification.
- Upload an epidemiological investigation report with more than 10 pages, check whether the parsing task time is within the value set for
PARSE_FILE_TIMEOUT_SECONDS, with no timeout errors. - After enabling
customPdfParse, upload a standardized PDF table document for disease control, verify that table cell content is fully extracted, with no missing or misaligned content. - Adjust
chunkSizeto the 800–1200 character range, upload an epidemic analysis report, check that chunking covers complete chapter content with no cross-chapter splitting.
The values given are common starting points and should be measured against the reader'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.