What the data for this category looks like
Civil aviation airport business-related data mainly comes from official operation manuals, flight schedule PDFs, check-in and security inspection guidance documents, luggage fee standard Excel files, terminal navigation maps, and similar sources. The update rhythm varies by data type: flight schedules are updated quarterly, operation manuals are updated annually or synchronized with major process adjustments, and temporary notice documents are released as needed. Document structures fall into three categories: long-form PDFs, structured spreadsheet files, and scattered text notices. Core fields include flight number, terminal number, gate number, baggage claim area number, and similar items. Units involved include hours, meters, passenger trips, and more.
What constraints these characteristics impose on the document parsing and chunking step
The presence of long-form PDF documents means chunk sizes cannot be too small. Too-small chunks will split the complete semantics of a single process. Structured spreadsheet files require special parsing logic. Without this logic, associated information within tables will be lost. Documents with different update frequencies correspond to different maintenance needs. Regularly updated flight schedule documents require that parsed chunks can be quickly synchronized for updates. Numbered fields in documents must retain context during chunking. Without this context, retrieval cannot link to corresponding processes. Additionally, some documents contain large numbers of embedded images. Extracting only text will lose critical visual information.
How to set the configuration
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Civil aviation airport business documents mostly consist of process descriptions and numbered content. This range preserves the complete semantics of a single process |
chunkSplitter | Split by paragraph + heading | Airport documents have clear chapter headings and process paragraphs. Splitting this way avoids cross-process chunking |
customPdfParse | Enable table and image recognition mode | Airport documents contain large numbers of check-in counter, fee standard tables and flight schedule images. Enabling this mode preserves structured information |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Some airport annual operation manuals can reach hundreds of MB per file. This upper limit covers most business document scenarios |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Large PDF documents take longer to parse. This duration prevents parsing from timing out and interrupting |
chunkSettingMode | Manual configuration | Adjustment of chunking rules is required for different document types such as PDFs and Excel files. Manual configuration allows more flexible adaptation to scenarios |
Three common mistakes
- Scenario: After uploading an airport business PDF, the interface does not display embedded flight schedule images. The cause is that the image extraction configuration for
customPdfParseis not enabled. The default parsing logic only extracts plain text content. - Scenario: Retrieved text chunks do not carry associated metadata such as flight numbers and gate numbers. The cause is that document metadata fields are not bound in the chunk configuration, and structured fields from airport documents are not mounted as additional associated information.
- Scenario: Uploading an Excel-format luggage fee standard document fails during parsing. The cause is that the parsing support switch for Excel files is not enabled, or the set
UPLOAD_FILE_MAX_SIZEis smaller than the actual file size.
How to confirm the configuration is correct
- Upload an airport business PDF that contains tables and embedded images. Check if parsed chunks retain table structures and image information.
- Upload an Excel-format luggage allowance standard document. Confirm that structured fields such as luggage allowance tiers and corresponding cabin classes can be extracted after parsing.
- Initiate a retrieval request. Confirm that returned chunks include pre-configured metadata fields such as flight number and terminal number.
- Upload an airport operation manual that exceeds the default file size limit. Confirm that the system prompt matches the configured
UPLOAD_FILE_MAX_SIZE.
The values provided here are common starting points. They should be evaluated against samples from the relevant use case.
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.