Reference only; verify against the target FastGPT release and deployment documentation.
What this category of data looks like
Port and shipping manifest data primarily comes from port operation scheduling systems, vessel berthing management systems, customs declaration systems, and on-site operation record terminals. Data updates are generated per voyage, work shift, or fixed daily schedule. Single manifest documents are mostly tabular files with fixed headers, including fields such as vessel name, voyage number, cargo loading and unloading volume, berthing and departure time, operation berth, fee details, and more. Corresponding units for fields include tons, TEU, hours, RMB yuan, and others. Some documents also include short text descriptions of vessel dynamics.
What constraints do these characteristics impose on the document parsing and chunking link
The multi-field structured nature of port manifests requires the parsing process to retain relevance between fields, and avoid breaking complete information for a single operation segment during splitting. High-frequency updated operation data requires the parsing service to have sufficient response time, to avoid interruptions during large file parsing. Multiple source document formats, including Excel tables, CSV export files, PDF operation reports, and structured data exported from databases, require parsing tools to support multiple file types. Fixed fields and units require retaining the original data format and meaning during chunking, to avoid unit confusion or field misalignment.
How to configure the settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Information for a single operation segment in a port manifest usually includes 3-5 core fields. 800–1200 characters can fully retain all content of a single voyage or single berth operation |
chunkSplitter | Split by structured documents | Port manifests are mostly tabular data with fixed headers. Splitting by structured format prevents fields from being split across records and retains field relevance |
customPdfParse | Enabled | Port operation PDF manifests often contain nested tables and fixed-format report content. Enabling this setting can accurately extract all fields within tables |
PARSE_FILE_TIMEOUT_SECONDS | 300–600 seconds | Large batch voyage manifest files have significant size, requiring sufficient time to complete full parsing and avoid mid-parsing timeouts |
UPLOAD_FILE_MAX_SIZE | 2000 MB | Batch-imported port voyage manifest compressed packages may reach large sizes, requiring adaptation to the file size limit for batch operations |
chunkSettingMode | Manual configuration | Port manifest formats are fixed and unified. Manual configuration can accurately match the parsing and chunking requirements of the business scenario |
Three common configuration mistakes
- Phenomenon: The system prompts that it cannot parse imported Excel port manifests, and only two columns of data can be recognized for CSV files. Cause:
customPdfParseor structured parsing configuration is not enabled, and the default parsing rules do not adapt to multi-column tabular manifest formats. - Phenomenon: Connection failure errors occur when deploying the parsing service locally, and logs show timeout or connection refused. Cause: The
PARSE_FILE_TIMEOUT_SECONDSparameter is not configured correctly, or the local service port is not open for external access. - Phenomenon: Mixed fields or single records split into multiple unrelated segments appear in the generated manifest aggregation report. Cause: The
chunkSizevalue does not match the character length of a single operation segment, or thechunkSplitterconfiguration is not set to split by structured documents.
How to verify that the configuration is correct
- Upload a single standard port manifest file, check the parsed chunked content, and confirm that each chunk contains complete information for a single operation segment.
- Import an Excel format manifest file, check that all preset fields are correctly recognized with no missing or misaligned entries.
- Submit a large manifest file, confirm that no parsing timeout errors occur, and view the parsing completion duration via backend logs.
- Adjust the
chunkSizeparameter and re-run parsing, compare chunking results under different configurations, and confirm that the chunk length meets business requirements.
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.