API Fetcher¶
Fetches data from healthcare-related APIs with optional presets for common FHIR and medical endpoints. Supports single or parallel request execution, simple bearer-token authentication, automatic response format detection (JSON/XML/text), basic retries, and status reporting.
Usage¶
Use this node to retrieve data from healthcare APIs in either single-request or batch/parallel mode. Select a preset endpoint for common resources (e.g., FHIR Patients), or choose custom to provide your own URL. Optionally add a bearer token and query parameters. In parallel mode, provide a JSON array of request configs to fan out multiple GET requests concurrently and aggregate the results.
Inputs¶
| Field | Required | Type | Description | Example |
|---|---|---|---|---|
| preset_endpoint | True | CHOICE | Select a predefined healthcare API endpoint or 'custom' to use your own API URL. | fhir_patients |
| api_url | True | STRING | The API URL to fetch data from. Used only when preset_endpoint is 'custom'. | https://hapi.fhir.org/baseR4/Patient |
| execution_mode | True | CHOICE | Choose 'single' for one request or 'parallel' to execute multiple requests concurrently. | single |
| max_workers | False | INT | Number of concurrent workers when execution_mode is 'parallel'. | 4 |
| batch_requests | False | STRING | A JSON array describing multiple GET requests for parallel mode. Each item may contain 'url' and optional 'params' (JSON string of query parameters). | [{"url": "https://hapi.fhir.org/baseR4/Patient", "params": "{\"_count\": 5}"}, {"url": "https://api.covidtracking.com/v1/us/current.json"}] |
| auth_token | False | STRING | Bearer token for Authorization header. Leave empty if not required by the API. | |
| query_params | False | STRING | Query parameters as a JSON object string. Applied to the request URL. In parallel mode, per-request 'params' in batch_requests take precedence. | {"name": "John", "_count": 10} |
| max_items | False | INT | Maximum number of items to keep when the response is a JSON array (single mode). | 10 |
Outputs¶
| Field | Type | Description | Example |
|---|---|---|---|
| response_data | STRING | The fetched and processed response data as a string. For JSON responses, formatted JSON is returned. In parallel mode, this is a JSON array of each request’s processed data. | {"resourceType": "Bundle", "entry": [...]} |
| status_message | STRING | Human-readable summary of the operation, including success counts in parallel mode. | Successfully fetched 10 items from API |
| success | BOOLEAN | Indicates if the operation succeeded. In parallel mode, true if at least one request succeeded. | True |
| status_code | INT | HTTP status code for single requests. In parallel mode, returns 200 for the aggregated result. | 200 |
| response_headers | STRING | Response headers as a JSON string. In parallel mode, contains an 'execution_log' array with per-request outcomes. | {"Content-Type": "application/fhir+json"} |
| response_time | FLOAT | Total elapsed time in seconds for the operation. | 1.42 |
Important Notes¶
- Preset endpoints include: fhir_patients, fhir_observations, fhir_medications, fhir_conditions, fhir_diagnostic_reports, covid_tracker, drug_interaction_checker, medical_codes_icd10, clinical_trials.
- api_url is only used when preset_endpoint is 'custom'.
- Only GET requests are supported in this simplified interface; use batch_requests for multiple GETs in parallel.
- Response format is auto-detected (JSON/XML/text). JSON arrays are truncated to max_items in single mode.
- Parallel mode aggregates results and sets success to true if any request succeeds; status_code is set to 200 for the aggregate.
- Default timeout per request is approximately 30 seconds with basic retry logic (up to 3 attempts).
- query_params and per-item 'params' in batch_requests must be valid JSON object strings (e.g., "{\"_count\": 10}").
- For protected APIs, provide auth_token as a bearer token. Never embed secrets directly in flows that may be shared.
- When handling patient or clinical data, ensure appropriate compliance and safeguards.
Troubleshooting¶
- Invalid batch_requests: Ensure it is a JSON array string. Example: "[{\"url\": \"https://...\"}]".
- Invalid query_params or params: Provide a valid JSON object string (e.g., "{\"_count\": 10}"), not URL-encoded or Python dict syntax.
- Authentication errors (401/403): Verify auth_token, endpoint requirements, and scopes. Some endpoints may not require authentication.
- Timeouts: Large responses or slow endpoints can exceed ~30s per request. Reduce max_items, narrow query parameters, or decrease batch size.
- Unexpected response format: If non-JSON is returned, the node outputs raw text or XML. Verify endpoint and parameters.
- Empty or partial results in parallel mode: Check the execution_log in response_headers for per-request messages and errors.
- Custom URL ignored: This occurs if preset_endpoint is not 'custom'. Set preset_endpoint='custom' to use api_url.