Upload Results¶

This guide shows how to create an analysis and persist result rows to the backend through ResultsAPI.

Create an Analysis¶

from owi.metadatabase.results import ResultsAPI
from owi.metadatabase.results.models import AnalysisDefinition
from owi.metadatabase.results.serializers import DjangoAnalysisSerializer

api = ResultsAPI(api_root="https://owimetadatabase-dev.azurewebsites.net/api/v1",
                 token="your-api-token")

definition = AnalysisDefinition(
    name="LifetimeDesignFrequencies",
    model_definition_id=12,
    location_id=None,
    source_type="notebook",
    source="scripts/data/results-example-data.xlsx",
    description="Lifetime design frequencies for Belwind.",
    additional_data={"sheet_name": "Lifetime -  Design frequencies"},
)

serializer = DjangoAnalysisSerializer()
payload = serializer.to_payload(definition)
created = api.create_analysis(payload)

analysis_id = int(created["id"])
print(f"Created analysis: {analysis_id}")

Upload Result Rows¶

Each result payload maps to a single Result row in the backend. ResultsAPI.create_results_bulk(...) renders a tqdm progress bar automatically.

from owi.metadatabase.results.serializers import DjangoResultSerializer

result_serializer = DjangoResultSerializer()

# Build payload from a ResultSeries object.
# analysis.to_results() returns a list of ResultSeries after validation.
payloads = [
    {**result_serializer.to_payload(series, analysis_id=analysis_id)}
    for series in result_series
]

# Bulk upload all payloads.
response = api.create_results_bulk(payloads)
print(f"Uploaded: {response['exists']}")

Upload a Single Row¶

If bulk upload is not supported by the target backend, the SDK can also create results one at a time:

from tqdm.auto import tqdm

for payload in tqdm(payloads, desc="Uploading result rows", unit="row"):
    api.create_result(payload)

Skipping Rows Without Resolved Locations¶

When working with location-scoped results, filter out rows where the turbine name could not be mapped to a backend location id:

upload_payloads = [p for p in payloads if p.get("location") is not None]
api.create_results_bulk(upload_payloads)

Create or Update Results in Bulk¶

When re-uploading data that may partially overlap with existing backend rows, use create_or_update_results_bulk. The method matches rows within the same analysis by their short_description and either creates or patches each one accordingly.

Every payload must include both analysis (the analysis id) and short_description keys:

payloads = [
    {"analysis": analysis_id, "short_description": "BBA01 - FA1", "name_col1": "frequency", ...},
    {"analysis": analysis_id, "short_description": "BBA02 - FA1", "name_col1": "frequency", ...},
]

result = api.create_or_update_results_bulk(payloads)
print(result["summary"])  # list of {analysis, short_description, result_id, action}

The returned summary list indicates the action taken for each row:

Action	Meaning
`created`	No matching row existed; a new row was inserted.
`updated`	An existing row with the same `short_description` was patched.

Ambiguous matches

If the backend already contains multiple rows with the same short_description for a given analysis, the call raises InvalidParameterError to avoid silent data corruption. Resolve the duplicates on the backend first, then retry.