async def get_model_content_handler(
job_id: UUID, uid: str = Depends(get_user())) -> JSONResponse:
"""API handler used to retrieve models
for a given user
Returns:
JSONResponse: [description]
"""
LOGGER.debug('retrieving models for user %s', uid)
# get all models from postgres database and convert to dict
job = get_user_job(PG_CREDENTIALS, uid, job_id)
if job is None:
LOGGER.error('unable to find job %s for user %s', job_id, uid)
return json_response_with_message(status.HTTP_404_NOT_FOUND,
'Cannot find specified job')
s3_data = retrieve_s3_file('/tensor-trigger/input-data' + str(job.job_id))
# generate metadata for file (including mime type) and convert to
# base64 encoded format
meta = Base64FileMetadata(file_size=0, mime_type='text/plain')
content = {
'http_code': status.HTTP_200_OK,
'job': generate_base64_file(s3_data, meta)
}
return JSONResponse(status_code=status.HTTP_200_OK, content=je(content))
async def get_models_handler(uid: str = Depends(get_user())) -> JSONResponse:
"""API handler used to retrieve models
for a given user
Returns:
JSONResponse: [description]
"""
LOGGER.debug('retrieving models for user %s', uid)
# get all models from postgres database and convert to dict
jobs = [j._asdict() for j in get_user_jobs(PG_CREDENTIALS, uid)]
content = {'http_code': status.HTTP_200_OK, 'jobs': jobs}
return JSONResponse(status_code=status.HTTP_200_OK, content=je(content))
async def get_model_meta_handler(model_id: UUID, uid: str = Depends(get_user())) -> JSONResponse:
"""API handler used to retrieve model by
model ID for a given user
Returns:
JSONResponse: [description]
"""
LOGGER.debug('retrieving model %s for user %s', model_id, uid)
model_meta = get_user_model(PG_CREDENTIALS, uid, model_id)
if model_meta is None:
return json_response_with_message(status.HTTP_404_NOT_FOUND, 'Cannot find specified model')
content = {'http_code': status.HTTP_200_OK,
'model': model_meta._asdict()}
return JSONResponse(status_code=status.HTTP_200_OK, content=je(content))
async def get_model_meta_handler(
job_id: UUID, uid: str = Depends(get_user())) -> JSONResponse:
"""API handler used to retrieve models
for a given user
Returns:
JSONResponse: [description]
"""
LOGGER.debug('retrieving models for user %s', uid)
# get all models from postgres database and convert to dict
job = get_user_job(PG_CREDENTIALS, uid, job_id)
if job is None:
LOGGER.error('unable to find job %s for user %s', job_id, uid)
return json_response_with_message(status.HTTP_404_NOT_FOUND,
'Cannot find specified job')
content = {'http_code': status.HTTP_200_OK, 'job': job._asdict()}
return JSONResponse(status_code=status.HTTP_200_OK, content=je(content))
async def get_model_handler(model_id: UUID, uid: str = Depends(get_user())) -> JSONResponse:
"""API handler used to retrieve model by
model ID for a given user
Returns:
JSONResponse: [description]
"""
LOGGER.debug('retrieving model %s for user %s', model_id, uid)
model_meta = get_user_model(PG_CREDENTIALS, uid, model_id)
if model_meta is None:
return json_response_with_message(status.HTTP_404_NOT_FOUND, 'Cannot find specified model')
s3_data = retrieve_s3_file('/tensor-trigger/' + str(model_id))
# generate metadata for file (including mime type) and convert to
# base64 encoded format
meta = Base64FileMetadata(file_size=0, mime_type='application/octet-stream')
content = {'http_code': status.HTTP_200_OK,
'model': generate_base64_file(s3_data, meta)}
return JSONResponse(status_code=status.HTTP_200_OK, content=je(content))
async def get_job_results_handler(
job_id: UUID, uid: str = Depends(get_user())) -> JSONResponse:
"""API handler used to retrieve models
for a given user
Returns:
JSONResponse: [description]
"""
LOGGER.debug('retrieving models for user %s', uid)
# get all models from postgres database and convert to dict
job = get_user_job(PG_CREDENTIALS, uid, job_id)
if job is None:
LOGGER.error('unable to find job %s for user %s', job_id, uid)
return json_response_with_message(status.HTTP_404_NOT_FOUND,
'Cannot find specified job')
if job.job_state != 2:
LOGGER.error(
'unable to retrieve job results for %s: invalid job state %s',
job_id, job.job_state)
return json_response_with_message(status.HTTP_400_BAD_REQUEST,
'Invalid job state')
meta = get_user_model(PG_CREDENTIALS, uid, job.model_id)
schema = meta.model_schema
s3_data = retrieve_s3_file('/tensor-trigger/output-data' + str(job.job_id))
results = json.loads(s3_data.getvalue().decode())
# generate metadata for file (including mime type) and convert to
# base64 encoded format
meta = Base64FileMetadata(file_size=0, mime_type='text/plain')
content = {
'http_code':
status.HTTP_200_OK,
'results':
_format_output_vector(results.get('output', []),
schema.get('output_schema'))
}
return JSONResponse(status_code=status.HTTP_200_OK, content=je(content))