What this category of data looks like
Data for this category primarily comes from the 12345 citizen hotline platform, contractor business systems, official knowledge bases, follow-up evaluation data, and local hotline management-related documents. Daily processed documents include request registration forms, dispatch records, handling responses, follow-up records, extension applications, return explanations, and knowledge base entries. Each single document typically contains fields such as request content, caller information, handling time limit, and responsible unit. Data is updated daily, covering full-process records from request receipt to closure.
Constraints imposed by these characteristics on the "document parsing and chunking" link
Decentralized data sources lead to diverse document formats, including structured business tables, unstructured response texts, and policy documents. Parsing logic must support multiple types. The daily update frequency requires efficient parsing processes to avoid ticket backlogs. Documents contain key business fields such as request type and responsible unit. Chunking must retain field associations to prevent key information from being split and broken. Some documents involve repeated requests or return records. Chunking must ensure accurate context association to support the accuracy of workflow guidance.
How to set the configurations
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Adapts to the length of request content in 12345 ticket documents, retains complete associations of key business fields, and avoids context breaks caused by excessive splitting |
chunkSplitter | Split by paragraph + semantic grouping | Adapts to the mixed structured and unstructured format of ticket documents. Prioritizes splitting by natural paragraphs, then merges short paragraphs based on semantics to ensure complete request details |
customPdfParse | Enabled | Adapts to partially exported PDF-format ticket documents, retains original layout and field layout, and improves parsing accuracy |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Adapts to batch-imported documents such as dispatch records and knowledge base entries, and prevents parsing interruptions caused by oversized files |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Adapts to batch parsing scenarios for multiple documents, and avoids timeout errors triggered by excessively long single parsing durations |
chunkSettingMode | Manual configuration | Supports custom chunking parameters for the differentiated formats of ticket documents, adapting to the needs of different types of ticket documents |
Three common mistakes
- Phenomenon: After enabling
customPdfParse, PDF-format dispatch records still cannot be parsed, and logs prompt parsing failure. Cause: The key information of the third-party PDF parsing tool is not correctly associated, or the corresponding parsing service address is not configured in the system. - Phenomenon: After uploading documents with multi-level directories, only the content of the top-level directory can be retrieved, and the content of subdirectories is missing. Cause: The multi-level directory recursive parsing setting is not enabled. Chunking only grabs the top-level document content and does not traverse subdirectory entries.
- Phenomenon: Call the API to submit ticket document parsing, the interface returns a successful status but no parsing result is returned. Cause: The API callback address is not correctly configured, or the result return logic is not triggered when the parsing task times out.
How to confirm the configuration is complete
- Upload a single request registration form document, view the parsed chunked content, confirm that key fields such as request type and responsible unit are not split and broken. Adjust corresponding parameters to values that match the current document format.
- Batch upload multiple ticket documents of different formats, check the completion status of parsing tasks, and confirm that the configured timeout parameters cover the parsing duration of most documents.
- Call the API to submit a test document, check whether the returned parsing result is consistent with the parsing result of manual upload, and confirm that the API parsing configuration is synchronized with the interface configuration.
- Check the key configuration page of the third-party parsing tool, confirm that the key corresponding to
customPdfParsehas been correctly entered into the system, with no null values or incorrect characters.
The values provided 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.