What This Category’s Data Looks Like
Oil and gas extraction due diligence data sources include official oil and gas field exploration and development reports, drilling site operation and maintenance logs, operating permit documents publicly released by industry regulators, and third-party oil and gas resource assessment documents. Update frequencies vary by data type: real-time drilling data updates as operations progress, annual exploration reports update per project cycle, and regulatory documents are released when policies or permits change. Documents mostly consist of structured tables paired with paragraph explanations. Each document includes fields such as well ID, operation block, reservoir burial depth, daily oil production per well, and operation cycle. Unit examples include meters, tons, cubic meters, and hours.
What Constraints These Characteristics Impose on the "Citation Sources and Traceability" Workflow
Dispersed data sources and inconsistent update frequencies require precise matching of document release times to corresponding operation nodes during traceability, to avoid citing outdated exploration or operation data. Many structured fields with clear associations mean traceability requires locating specific well IDs or operation paragraphs, not entire documents, and retaining associated identifiers like well numbers and operation batches. Individual documents can be lengthy, some containing multiple sets of well data. Traceability requires avoiding splitting cross-well associated parameters, and distinguishing regulatory document identifiers from different sources to avoid confusing documents with the same name.
How to Configure Settings
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
top_k | 8-12 | Oil and gas extraction due diligence documents are lengthy and contain multiple sets of well data. This value range covers core operation parameter sources while avoiding redundant context |
similarity_threshold | 0.72-0.85 | Domain terminology is highly specialized. This interval filters irrelevant documents while retaining associated operation and maintenance logs and exploration reports |
chunk_size | 800-1200 characters | Individual documents include associated fields such as well numbers and reservoir parameters. This segment length preserves complete associated information within paragraphs |
rerank_top_k | 4-6 | Prioritize the most relevant core due diligence data sources, avoiding confusion from too many reranked results disrupting traceability information |
enable_citation | Enabled | Knowledge base identifiers, document titles, and specific paragraph positions must be displayed in outputs to meet traceability requirements |
max_context_tokens | 12000-15000 | Citations for oil and gas extraction due diligence reports are lengthy. This value range fully preserves contextual information associated with traceability |
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, and it is recommended to test on your own samples before finalizing settings.
Three Common Misconfigurations
- Issue: When calling the chat interface, the returned
source_infofield is empty or does not include the knowledge base ID. Cause: Theenable_citationconfiguration item is not enabled, or the interface request does not carry the corresponding parameters for enabling traceability. - Issue: Returned citation sources merge paragraphs from multiple documents, making it impossible to locate the specific well ID or operation link of a single document. Cause: The
chunk_sizesetting exceeds the paragraph length of a single document, causing associated cross-well data to be split during parsing. - Issue: The returned reference file does not match the knowledge base content actually called, with incorrect traceability references. Cause: The
similarity_thresholdis set too low, introducing irrelevant non-target knowledge base documents, or unique well ID or operation batch identifiers are not added to knowledge base documents.
How to Confirm Configurations Are Set Correctly
- Initiate a test query containing oil and gas extraction domain terminology, and check if the returned results include information about knowledge base identifiers, document titles, and specific paragraph positions.
- Review the segmented content parsed by the knowledge base, confirm that segments do not span multiple sets of well data or operation data, and align with the
chunk_sizesetting logic. - Adjust the values of
top_kandrerank_top_k, and verify that the number of returned citation sources matches the expected configuration. - Call the interface and parse the returned
source_infofield, confirm that it contains complete traceability associated information, with no missing or incorrect identifiers.
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.