What Data for This Category Looks Like
Photovoltaic intelligent due diligence reports are core documents used by financial institutions for credit granting and underwriting of photovoltaic projects. Data sources include on-site project survey records, grid operation ledgers, factory quality inspection documents from component manufacturers, irradiance data from local meteorological stations, and compliance filing documents.
Update frequency varies by project phase. Survey data is updated weekly during the project initiation phase. Power generation and operation data is synced monthly after grid connection and official operation.
Document structure includes modules such as project location description, component peak power parameters, annual irradiance statistics, operation logs, and grid connection permits. Fields include unique identifiers like component peak power (unit: Wp), annual irradiance (unit: kWh/㎡), filing number, and grid connection date. A single report often includes attachments in multiple formats.
What Constraints Do These Characteristics Impose on Deployment and Upgrade
The multi-source, multi-format data characteristics of photovoltaic due diligence reports require the deployment process to support mixed file parsing for CAD, PDF, and Excel files. Default parsing timeout configurations must be adjusted to accommodate large attachments.
The high data update frequency requires configuring scheduled synchronization and incremental update mechanisms. This avoids resource waste caused by full knowledge base rebuilds.
The professional and diverse field requirements demand support for custom field mapping. This ensures the parsed knowledge base accurately matches the exclusive parameters used in due diligence queries.
The upgrade process must balance knowledge base sharing across multiple instances. This prevents data inconsistency caused by independent deployments.
How to Set Configuration
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
PARSE_FILE_TIMEOUT_SECONDS | 900 seconds | Photovoltaic due diligence reports often include CAD drawings, multi-page PDFs, and multiple Excel ledgers. Single-file parsing takes significant time |
UPLOAD_FILE_MAX_SIZE | 2000 MB | A single large photovoltaic project due diligence report may include multiple attachments. Large file upload support is required |
maxContext | 8000–12000 characters | Photovoltaic due diligence reports have dense fields. Sufficient context is needed to retain complete project parameters and compliance information |
RECALL_COUNT | Top 10 entries | Enough project details must be retrieved to cover multi-dimensional information required for due diligence |
SIMILARITY_THRESHOLD | 0.75–0.85 | Filter low-relevance meteorological or operation data, and retain highly matched core project parameters |
RERANK_TOP_N | Top 5 entries | Focus on the most relevant core data to avoid redundant information interfering with due diligence conclusions |
The parameter values provided on this page are common starting points for configuration. Actual values are affected by material format, data volume, and business rules. Specific issues require specific analysis. It is recommended to test on your own samples before finalizing settings.
Three Common Misconfigurations
- Parsing timeout after uploading a large photovoltaic due diligence report, with
ETIMEDOUTerrors shown in logs. ThePARSE_FILE_TIMEOUT_SECONDSparameter is not adjusted, and the default value is insufficient for parsing large documents with multiple attachments. - 401 unauthorized errors returned when calling the reranking model. The model API key is not configured correctly, or the used image is not bound to valid access credentials. This does not match the reranking model configuration required for photovoltaic due diligence.
- Knowledge base synchronization data inconsistency after deploying two FastGPT instances. The official recommended multi-instance knowledge base sharing solution is not used. Using external API connections causes data synchronization delays or permission conflicts.
How to Confirm Configuration Is Correct
- Upload a standard photovoltaic project due diligence report. Check if parsing completes within the preset timeout period.
- Initiate a due diligence query for project parameters. Check that the reranking model call logs show no 401 errors.
- Manually trigger a knowledge base incremental synchronization. Check if new project data is synced to the knowledge base within the preset interval.
- View the system parameter panel. Confirm that the
UPLOAD_FILE_MAX_SIZEsetting matches the local storage quota.
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.