What the data for this category looks like
Data sources include public industry research reports, public grid planning documents, publicly disclosed materials from energy storage enterprises, and broker research reports. Update frequency varies by document type: industry reports are updated quarterly, policy documents are adjusted irregularly alongside policy releases, and enterprise operational data is updated along with financial reporting cycles. Most documents are long-form PDFs, some with embedded charts and tables. Content covers energy storage installed capacity scale, system cycle parameters, cost accounting, grid connection specifications, and similar topics. Units include GW, MWh, yuan/kWh, and some documents have mixed unit usage.
What constraints do these characteristics impose on the "document parsing and chunking" workflow?
Long-form documents can result in chunks that are either too long to fit within context windows, or too short to retain professional contextual links. Chunk granularity must be adjusted to match the document structure. Embedded charts and tables require the parsing module to support structured extraction of table content, rather than only extracting scattered text. Mixed unit usage can cause data association failures after chunking, so unit context must be retained during the parsing stage. The pre-OCR step for scanned PDFs increases parsing time, so timeout thresholds need to be adjusted. Paragraphs dense with professional terminology require avoiding chunking across terms, to ensure semantic completeness within individual chunks.
How to configure the settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
chunk_size | 800–1200 characters | Energy storage research reports often contain professional parameters and long paragraphs. This range balances context completeness and retrieval accuracy |
chunk_overlap | 100–150 characters | Prevents professional terms or parameters from being truncated during chunking, and ensures cross-chunk semantic association |
enable_table_parse | Enabled | Energy storage research reports contain large amounts of tabular data such as installed capacity and costs. Enabling this option allows structured extraction of table content |
PARSE_FILE_TIMEOUT_SECONDS | 300 seconds | Long documents and scanned PDFs have longer OCR and parsing times. This value covers most scenarios |
UPLOAD_FILE_MAX_SIZE | 1000 MB | Most energy storage research reports are long PDFs with multiple chapters per single file. This value supports large file uploads |
ocr_enable | Enabled based on document type | Scanned PDFs require OCR to be enabled. Plain text PDFs can disable OCR to improve parsing speed |
The parameter values provided on this page are common recommended starting points for configuration. Actual values are affected by material format, data volume, and business rules. Specific issues require individual analysis. It is recommended to test on your own samples before finalizing settings.
Three common configuration mistakes
- Scenario: When uploading a single energy storage research report PDF that exceeds the configured threshold, an "offset out of range" error appears when progress reaches approximately 90%. Cause: The
UPLOAD_FILE_MAX_SIZEparameter is not set correctly, or the chunked upload mechanism is not enabled during upload. This causes index misalignment after large file transfer interruption. - Scenario: Parsed chunked content lacks energy storage-related tabular parameters such as installed capacity and cycle times. Cause: The
enable_table_parseconfiguration is not enabled, only plain text content is extracted, and embedded tables are not parsed structurally. - Scenario: Chunking results for energy storage research reports uploaded via API are inconsistent with those uploaded directly via the platform. Cause: The
chunk_sizeandchunk_overlapparameters are not specified during API calls, using default configurations that do not match those used for manual platform uploads, or the document parsing mode is not set correctly.
How to confirm configurations are properly applied
- Upload a typical energy storage research report PDF, view the parsed chunk list, and verify that chunk lengths fall within the expected range.
- View the tabular content in the parsing results, confirm that energy storage-related parameters and unit information are fully extracted.
- Upload the same document via both the platform and API, compare the granularity and content of the chunking results, and confirm configuration consistency.
- Upload a scanned energy storage research report, check whether the text after OCR recognition is complete, and confirm that the
ocr_enableconfiguration is active.
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.