What the data for this category looks like
Data related to energy storage marketing mainly comes from product technical manuals, industry compliance certification documents, scenario adaptation plans, marketing script packages, and grid operation case documents. Updates occur when new products launch, industry policies adjust, or operation cases are updated, with no fixed cycle. Document structures include structured parameter tables, scenario-based descriptive text, compliance mark explanations, and version annotation information. Common fields include energy storage system capacity (units: kWh or MWh), cycle life (units: cycles), operating temperature range (units: ℃), and certification marks.
What constraints these characteristics impose on document parsing and chunking
The presence of structured parameter tables requires the parsing process to retain the row and column correspondence of cells, to avoid misalignment of parameter content. Scenario-based descriptive text and technical parameters often have strong correlations. Chunking must avoid splitting related content into different chunks. Scanned marketing materials and older version manuals account for a high proportion. OCR must be supported to recognize text in images. Product versions are updated frequently. Chunking must include metadata to distinguish content from different versions. Excel-format parameter documents are common. Table structure parsing must be supported; simple text extraction cannot meet parsing needs.
How to set configurations
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
maxChunkSize | 800–1200 characters | Energy storage documents contain long sections of technical parameters and scenario descriptions. This range balances parameter completeness and semantic coherence |
chunkOverlap | 100–150 characters | Prevents technically related parameters across chunks from being split apart, and ensures semantic association between related fields such as cycle life and capacity |
enableOCR | Enabled | Energy storage marketing materials often include scanned product brochures and manuals. OCR is required to recognize printed text content |
parseExcelStruct | Strictly retain cell mapping | Most energy storage parameters are organized in Excel tables. Retaining structure prevents misalignment of parameter correspondences |
extractMetadata | Enabled, includes version number and release date | Energy storage products are updated frequently. Metadata can be used for chunk traceability and compliance verification |
PARSE_TIMEOUT | 300 seconds | Large energy storage industry white papers take a long time to parse. This duration covers the complete parsing process |
The parameter values provided on this page are common starting points for configuration. Actual values are affected by material form, data volume, and business rules. Specific issues require specific analysis. It is recommended to test on your own samples before finalizing settings.
Three common errors
- Phenomenon: Scanned energy storage manual uploads result in empty parsing results or only a small amount of extracted text. Cause: The
enableOCRconfiguration is not enabled, so text in scanned images cannot be recognized. - Phenomenon: After uploading an Excel-format energy storage parameter table, the vectorized index results have misaligned fields. Cause: The
parseExcelStructconfiguration is not enabled. The default parsing breaks table content into unstructured text. - Phenomenon: After calling the knowledge base upload API, the specified chunk content is not correctly added. Cause: The format of the
customChunkContentparameter is not set correctly, or it does not match the chunking rules after document parsing.
How to verify the configuration is properly set
- Upload a single scanned energy storage product manual, check if the parsing result includes complete printed text, to confirm the
enableOCRconfiguration is active. - Upload an energy storage parameter Excel table, verify that the parsed chunks retain the row and column correspondence of cells, to confirm the
parseExcelStructconfiguration is active. - Call the knowledge base upload API, pass custom chunk content, check if the content is included in the knowledge base chunk list, to confirm the API parameter configuration is correct.
- Check the parsing log, confirm that the values of
maxChunkSizeandchunkOverlapare applied, and the chunk length matches the preset range.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-14.