What Data for This Category Looks Like
Data for this category primarily comes from formula systems, raw material batch ledgers, internal mixers, Mooney viscometers, vulcanizing presses, and laboratory testing equipment. Daily documents include formula sheets, weighing records, mixing curves, rubber compound quick test reports, vulcanization process cards, mold usage records, and finished product performance test reports. Corresponding records are generated for each truckload of raw materials, each batch of rubber compound, and each mold production run. Most document fields include batch numbers, process parameters, test values, with specific units such as degrees Celsius, minutes, kilograms, and others.
Constraints on Document Parsing and Chunking
The high-frequency, small-batch update nature of this category requires document parsing and chunking to adapt to multiple frequently generated document formats, and avoid breaking cross-document association logic via chunking. Structural differences between document types are significant. Formula sheets contain structured feeding parameters. Mixing curves are chart documents with time-series data. Vulcanization process cards include fixed processes and dynamic parameters. Chunking must retain the binding relationship between fields and units, otherwise corresponding process conditions cannot be matched during traceability. Processes in this category are irreversible. Incorrect parameter associations cannot be corrected through secondary processing. The timeliness requirements for single-batch data mean the parsing link cannot have excessive delays, otherwise real-time defect analysis cannot be supported.
Configuration Settings
This guide is written for FastGPT v4.17.0:
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Adapts to the content length of documents in this category that contain both structured parameters and process descriptions, avoiding splitting that breaks the binding relationship between parameters and units |
chunkSplitter | By paragraph and semantic boundary | Adapts to documents with time-series or process flows such as mixing curves and process cards, retains contextual logic, and avoids splitting that breaks process associations |
customPdfParse | Enable chart recognition mode | Adapts to PDF documents with visualized data such as mixing curves, extracts time-series parameters and corresponding values within charts |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Addresses the parsing time of PDF documents containing large amounts of data such as mixing curves, avoids parsing task failures caused by timeouts |
UPLOAD_FILE_MAX_SIZE | 100 MB | Adapts to batch-imported multi-batch production record documents, meets the capacity requirements for single uploads |
chunkSettingMode | Manual configuration | Adjusts chunking rules for different document types, matches structural differences between formula sheets and record-style documents |
Three Common Configuration Errors
Unsupported Format or Missing Fields
Phenomenon: An error indicating unsupported format is prompted when uploading Excel-format formula sheets or mixing records, or fields are missing after import. Cause: The parsing configuration adapted to structured tables is not enabled. The default parsing rules only support CSV formats with a small number of columns, and cannot recognize multi-column structured data in Excel.
Missing Chart Parameters in Parsed Mixing Curve PDFs
Phenomenon: When parsing a mixing curve PDF, the returned text content lacks chart parameters, and only page text is retained. Cause: The chart recognition mode of customPdfParse is not enabled. The default parsing only extracts plain text content, losing key process values in time-series curves.
Batch Import Parsing Timeouts
Phenomenon: Parsing timeouts occur when batch importing multi-batch production records, and the task status shows failure. Cause: The value of PARSE_FILE_TIMEOUT_SECONDS is not adjusted. The default timeout duration is insufficient for processing mixing curve documents containing large amounts of data.
How to Confirm Proper Configuration
- Upload a single mixing curve PDF document, check if the parsing result includes time-series parameters and corresponding values within the chart, confirming that the chart recognition configuration is active.
- Import multi-format documents (Excel, CSV, PDF), verify that fields from different types of documents are fully extracted, with no format unsupported errors.
- Batch upload batch production records, monitor the execution duration of parsing tasks, confirm that no timeout failures occur, and that the configured timeout value adapts to the current document scale.
- Review chunked document fragments, confirm that parameters and units are not split into different chunks, and that chunk length and segmentation rules match the document structure.
The values provided here are common starting points, and should be evaluated against the reader's own samples.