What this category of data looks like
Data sources include the national unified declaration and approval system for environmental impact assessment (EIA) of construction projects, ecological environment zoning control results, pollution discharge permit platform, and monitoring data. Update frequency is irregular. Daily processed documents include environmental impact assessment reports or report forms, approval applications, public participation explanations, technical evaluation opinions, and EIA approvals. These documents have standardized chapter structures, with dedicated fields such as project code, EIA level, pollution prevention measures, and public participation status. Some documents have complex formats and long multi-page lengths.
Constraints for Document Parsing and Chunking
Multiple data sources and diverse file formats require parsing functions to support PDF, Word, Excel, and other formats. Standardized document structures but wide variation in length require chunking rules to identify chapter boundaries, and avoid splitting key content across chapters. Irregular updates and severe error consequences require chunking to preserve core business information, while adapting to documents of all sizes. Differences in data format across source systems require the parsing link to support field extraction from varied formats.
Configuration Settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunkSize | 800–1200 characters | Matches the length of core content per chapter in EIA documents, avoids splitting key business information |
chunkSplitter | Split by chapter titles | Aligns with the standardized chapter structure of EIA documents, preserves the integrity of business content per chapter |
customPdfParse | Enabled | Adapts to most EIA documents being scanned or formatted PDFs, extracts complete text content |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds | Follows the default threshold for large document parsing in FastGPT v4.17.0, accommodates parsing time for large EIA reports |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Accommodates the file size of a single large EIA report |
chunkSettingMode | Custom mode | Allows adjustment of chunking rules based on the characteristics of EIA documents, adapts to dedicated business scenarios |
Three Common Misconfigurations
- Phenomenon: No matching results appear after uploading an EIA PDF file, and the interface prompts a parsing error. Cause: The
customPdfParseconfiguration is not enabled, so text content in the document cannot be fully extracted. - Phenomenon: Chunk results include unrelated content across chapters, such as combining public participation explanations and pollution prevention measures into one chunk. Cause: The
chunkSplitteris not set to split by chapter titles, and the general character-based chunking rule is used instead. - Phenomenon: Related configuration options for custom PDF parsing are not displayed in the interface. Cause: The
chunkSettingModeis not switched to custom mode, so advanced parsing configurations are not shown.
How to Confirm Proper Configuration
- Upload a small EIA report form file, and verify that the parsed text content is complete, with no obvious truncation or garbled characters.
- Navigate to the knowledge base search test function, enter "pollution prevention measures for a project", and confirm that returned results come from chunked content of the corresponding chapter.
- Check the
PARSE_FILE_TIMEOUT_SECONDSconfiguration value, and confirm it is set to 600 seconds or higher to accommodate large document parsing needs. - Review the uploaded file size limit configuration, and confirm
UPLOAD_FILE_MAX_SIZEis set to no less than 1000 MB to support large EIA reports.
The values provided are common starting points and should be evaluated 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.