Data Characteristics
Infectious disease product data comes primarily from clinical research reports, drug inserts, diagnostic reagent inserts, academic journal articles, and data published by national Centers for Disease Control and Prevention (CDC). This data updates frequently, especially for new pathogens or antibiotic resistance surveillance, with updates potentially weekly or monthly. Document structures typically include fixed sections like disease overview, etiology, epidemiology, clinical manifestations, diagnostic methods, treatment plans, prognosis, and prevention measures. Fields often include pathogen name, host, infection site, transmission route, incubation period, symptoms, diagnostic markers, reagent sensitivity and specificity, drug mechanism of action, dosage, adverse reactions, and resistance genes. Units are often international standard units, such as mg/kg, IU/mL, copies/mL, as well as percentages and multiples.
Constraints from "Forms and Interactions"
The high update frequency of infectious disease data requires the knowledge base to support rapid synchronization and incremental updates. This prevents users from receiving outdated information. The complexity of document structures and the diversity of fields necessitate form designs that balance generality and specialization. This ensures accurate capture of user queries regarding specific pathogens, drugs, or diagnostic reagents. For example, a user might need to query the resistance spectrum of a pathogen or the diagnostic performance of a specific diagnostic reagent for an infectious disease. The specialized nature of fields and the strictness of units demand high accuracy for interactive Q&A. The system must understand and correctly parse professional terminology and values entered by users, then return results in a standardized format. This includes recognizing aliases and abbreviations, and evaluating numerical ranges or critical values.
Configuration Settings
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
maxContext | 8192 | Ensures the capacity for lengthy questions and context, including detailed clinical symptoms, diagnostic standards, and treatment plans. |
Chunk size (Chunk Length) | 800–1200 characters (characters) | Balances semantic completeness of long clinical documents with recall efficiency, preventing information fragmentation from excessive splitting. |
Recall count (Recall Count) | Top 10 entries (top 10) | Infectious disease information is highly interconnected. Increasing recall count improves coverage and reduces the risk of missing critical information. |
Similarity threshold (Similarity Threshold) | 0.78–0.85 | Ensures professional relevance of recalled content, filtering out noise, especially for queries about disease symptoms and diagnostic standards. |
Rerank result count (Reranked Return Count) | Top 5 entries (top 5) | Based on a high recall count, reranking focuses on the most relevant content, improving efficiency for users to obtain core information. |
PARSE_FILE_TIMEOUT_SECONDS | 600 seconds (seconds) | Accommodates parsing time for large clinical guidelines or research reports, preventing parsing timeouts due to excessively large files. |
Common Pitfalls
- Chat dialog displays "knowledge base not selected," but debug preview is normal: This usually occurs because the knowledge base reference in the workflow is not correctly bound to the session context, preventing the knowledge base ID from being obtained at runtime.
- AI model dropdown for question classification in the workflow is empty: This is typically due to incorrect AI model service configuration or failed model interface calls, preventing the loading of available model lists.
- Returned reference content displays even when references are disabled: This phenomenon may relate to default behavior or configuration caching in specific versions (e.g.,
4.9.4). Check system-level or workflow-level reference display settings.
Verification Steps
- Submit complex queries containing professional terminology, disease names, and diagnostic reagent models. Observe whether the returned results include accurate professional information and reference sources.
- Simulate inquiries with numerical ranges and dosage units. Verify that the system returns correct values and units, and check if it correctly handles critical value queries.
- Immediately perform relevant queries after a knowledge base update. Verify that the system quickly reflects the latest data, confirming that the update mechanism functions correctly.
Note: The values provided are common starting points and should be measured 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 2026-09-21.