AntiNex Client Docs¶
AntiNex Stack Status¶
AntiNex client is part of the AntiNex stack:
| Component | Build | Docs Link | Docs Build |
|---|---|---|---|
| REST API | 
|
Docs | |
| Core Worker | 
|
Docs | |
| Network Pipeline | 
|
Docs | |
| AI Utils | 
|
Docs | |
| Client | 
|
Docs |
Table of Contents¶
These are the docs for the AntiNex client repository.
AI Client Class¶
This is the AntiNex Python Client class
-
class
antinex_client.ai_client.AIClient(user='user-not-set', password='password-not-set', url='http://localhost:8010', email='email-not-set', verbose=True, ca_dir=None, cert_file=None, key_file=None, debug=False)[source]¶ AntiNex Python AI Client
This can use either environment variables or keyword arguments to create a valid client.
-
build_response(status=3, error='', data=None)[source]¶ Parameters: - status – status code
- error – error message
- data – dictionary to send back
-
Building a Client from Environment Variables¶
This is how to build an AntiNex Python Client object from just environment variables.
-
antinex_client.build_ai_client_from_env.build_ai_client_from_env(verbose=True, debug=False, ca_dir=None, cert_file=None, key_file=None)[source]¶ Use environment variables to build a client
Parameters: - verbose – verbose logging
- debug – debug internal client calls
- ca_dir – optional path to CA bundle dir
- cert_file – optional path to x509 ssl cert file
- key_file – optional path to x509 ssl key file
AntiNex Consts¶
Here are the environment variables and constants used by the AntiNex client.
LOGIN_SUCCESS = 0
LOGIN_NOT_ATTEMPTED = 1
LOGIN_FAILED = 2
SUCCESS = 0
FAILED = 1
ERROR = 2
NOT_SET = 3
DISABLED = 4
Client Environment Variables¶
These are the environment variables used by the AntiNex client.
Note
Please make sure these match up with your local, running stack:
AntiNex REST API flags, endpoint and credentials https://github.com/jay-johnson/train-ai-with-django-swagger-jwt
AntiNex is running using compose.yml file: https://github.com/jay-johnson/train-ai-with-django-swagger-jwt/blob/master/compose.yml
ANTINEX_PUBLISH_ENABLED = bool(ev(
"ANTINEX_PUBLISH_ENABLED",
"1") == "1")
ANTINEX_URL = ev(
"ANTINEX_URL",
"http://localhost:8010")
ANTINEX_CA_FILE = os.getenv(
"ANTINEX_CA_FILE",
None)
ANTINEX_CERT_FILE = os.getenv(
"ANTINEX_CERT_FILE",
None)
ANTINEX_KEY_FILE = os.getenv(
"ANTINEX_KEY_FILE",
None)
ANTINEX_USER = ev(
"ANTINEX_USER",
"root")
ANTINEX_EMAIL = ev(
"ANTINEX_EMAIL",
"notreal@test.com")
ANTINEX_PASSWORD = ev(
"ANTINEX_PASSWORD",
"123321")
# provide a template request publish file like:
# https://github.com/jay-johnson/antinex-client/blob/master/examples/predict-rows-scaler-full-django.json
ANTINEX_PUBLISH_REQUEST_FILE = ev(
"ANTINEX_PUBLISH_REQUEST_FILE",
("/opt/antinex/client/examples/"
"predict-rows-scaler-full-django.json"))
# comma-separated list
ANTINEX_FEATURES_TO_PROCESS_STR = os.getenv(
"ANTINEX_FEATURES_TO_PROCESS",
None)
# comma-separated list
ANTINEX_IGNORE_FEATURES_STR = os.getenv(
"ANTINEX_IGNORE_FEATURES",
None)
# comma-separated list
ANTINEX_SORT_VALUES_STR = os.getenv(
"ANTINEX_SORT_VALUES",
None)
# comma-separated list
ANTINEX_METRICS_STR = os.getenv(
"ANTINEX_METRICS",
None)
# comma-separated list
ANTINEX_HISTORIES_STR = os.getenv(
"ANTINEX_HISTORIES",
None)
ANTINEX_ML_TYPE = ev(
"ANTINEX_ML_TYPE",
"classification")
ANTINEX_USE_MODEL_NAME = ev(
"ANTINEX_USE_MODEL_NAME",
"Full-Django-AntiNex-Simple-Scaler-DNN")
ANTINEX_PREDICT_FEATURE = ev(
"ANTINEX_PREDICT_FEATURE",
"label_value")
ANTINEX_SEED = int(ev(
"ANTINEX_SEED",
"42"))
ANTINEX_TEST_SIZE = float(ev(
"ANTINEX_TEST_SIZE",
"0.2"))
ANTINEX_BATCH_SIZE = int(ev(
"ANTINEX_BATCH_SIZE",
"32"))
ANTINEX_EPOCHS = int(ev(
"ANTINEX_EPOCHS",
"15"))
ANTINEX_NUM_SPLITS = int(ev(
"ANTINEX_NUM_SPLITS",
"3"))
ANTINEX_LOSS = ev(
"ANTINEX_LOSS",
"binary_crossentropy")
ANTINEX_OPTIMIZER = ev(
"ANTINEX_OPTIMIZER",
"adam")
ANTINEX_VERSION = ev(
"ANTINEX_VERSION",
"1")
ANTINEX_CONVERT_DATA = bool(ev(
"ANTINEX_CONVERT_DATA",
"1") == "1")
ANTINEX_CONVERT_DATA_TYPE = ev(
"ANTINEX_CONVERT_DATA_TYPE",
"float")
ANTINEX_INCLUDE_FAILED_CONVERSIONS = bool(ev(
"ANTINEX_INCLUDE_FAILED_CONVERSIONS",
"0") == "1")
ANTINEX_MISSING_VALUE = ev(
"ANTINEX_MISSING_VALUE",
"-1.0")
ANTINEX_PUBLISH_TO_CORE = bool(ev(
"ANTINEX_PUBLISH_TO_CORE",
"1") == "1")
ANTINEX_CHECK_MISSING_PREDICT = bool(ev(
"ANTINEX_CHECK_MISSING_PREDICT",
"1") == "1")
ANTINEX_CLIENT_VERBOSE = bool(ev(
"ANTINEX_CLIENT_VERBOSE",
"1") == "1")
ANTINEX_CLIENT_DEBUG = bool(ev(
"ANTINEX_CLIENT_DEBUG",
"0") == "1")
# set empty defaults
ANTINEX_FEATURES_TO_PROCESS = []
ANTINEX_IGNORE_FEATURES = []
ANTINEX_SORT_VALUES = []
ANTINEX_METRICS = []
ANTINEX_HISTORIES = []
These environment variables are set as lists based of commas:
if ANTINEX_FEATURES_TO_PROCESS_STR:
ANTINEX_FEATURES_TO_PROCESS = \
ANTINEX_FEATURES_TO_PROCESS_STR.split(",")
if ANTINEX_IGNORE_FEATURES_STR:
ANTINEX_IGNORE_FEATURES = \
ANTINEX_IGNORE_FEATURES_STR.split(",")
if ANTINEX_SORT_VALUES_STR:
ANTINEX_SORT_VALUES = \
ANTINEX_SORT_VALUES_STR.split(",")
if ANTINEX_METRICS_STR:
ANTINEX_METRICS = \
ANTINEX_METRICS_STR.split(",")
if ANTINEX_HISTORIES_STR:
ANTINEX_HISTORIES = \
ANTINEX_HISTORIES_STR.split(",")
Environment variables that are set to dictionaries for faster lookups:
FILTER_FEATURES_DICT = {}
FILTER_FEATURES = []
for idx, f in enumerate(ANTINEX_FEATURES_TO_PROCESS):
include_feature = True
if f == ANTINEX_PREDICT_FEATURE:
include_feature = False
for i in ANTINEX_IGNORE_FEATURES:
if f == i:
include_feature = False
break
if include_feature:
FILTER_FEATURES.append(f)
FILTER_FEATURES_DICT[f] = idx
# end of for all features not being ignored
Generate an AntiNex AI Request¶
This method will use the environment variables from the consts.py:
generate_ai_request Method¶
-
antinex_client.generate_ai_request.generate_ai_request(predict_rows, req_dict=None, req_file='/opt/antinex/client/examples/predict-rows-scaler-full-django.json', features=[], ignore_features=[], sort_values=[], ml_type='classification', use_model_name='Full-Django-AntiNex-Simple-Scaler-DNN', predict_feature='label_value', seed=42, test_size=0.2, batch_size=32, epochs=15, num_splits=3, loss='binary_crossentropy', optimizer='adam', metrics=[], histories=[], filter_features_dict={}, filter_features=[], convert_enabled=True, convert_to_type='float', include_failed_conversions=False, value_for_missing='-1.0', version='1', publish_to_core=True, check_missing_predict_feature=True, debug=False)[source]¶ Parameters: - predict_rows – list of predict rows to build into the request
- req_dict – request dictionary to update - for long-running clients
- req_file – file holding a request dict to update - one-off tests
- features – features to process in the data
- ignore_features – features to ignore in the data (non-numerics)
- sort_values – optional - order rows for scaler normalization
- ml_type – machine learning type - classification/regression
- use_model_name – use a pre-trained model by name
- predict_feature – predict the values of this feature
- seed – seed for randomness reproducability
- test_size – split train/test data
- batch_size – batch size for processing
- epochs – test epochs
- num_splits – test splits for cross validation
- loss – loss function
- optimizer – optimizer
- metrics – metrics to apply
- histories – historical values to test
- filter_features_dict – dictionary of features to use
- filter_features – list of features to use
- convert_to_type – convert predict_row values to scaler-ready values
- include_failed_conversions – should the predict rows include fails
- value_for_missing – set this value to any columns that are missing
- version – version of the API request
- publish_to_core – want to publish it to the core or the worker
- debug – log debug messages
AntiNex Client Utils¶
Utility methods
-
antinex_client.utils.convert_to_date(value=None, format='%Y-%m-%d %H:%M:%S')[source]¶ param: value - datetime object param: format - string format
Make Predictions with a Pre-trained Deep Neural Network¶
This uses environment variables to build, train and make predictions using the AntiNex client. If the deep neural network already exists it will use it to make new predictions. If it does not exist it will train a new one.
This python script is available in the pip: ai_env_predict.py
It takes parameters:
parser = argparse.ArgumentParser(
description=("Python client to make Predictions "
"using a Pre-trained Deep Neural Network "
"with AntiNex Django Rest Framework"))
parser.add_argument(
"-f",
help=("file to use default ./examples/"
"predict-rows-scaler-full-django.json"),
required=False,
dest="datafile")
parser.add_argument(
"-m",
help="send mock data",
required=False,
dest="use_fake_rows",
action="store_true")
parser.add_argument(
"-s",
help="silent",
required=False,
dest="silent",
action="store_true")
parser.add_argument(
"-d",
help="debug",
required=False,
dest="debug",
action="store_true")
Source Code¶
-
antinex_client.scripts.ai_env_predict.start_predictions()[source]¶ Using environment variables, create an AntiNex AI Client. You can also use command line args if you want.
This can train a new deep neural network if it does not exist or it can use an existing pre-trained deep neural network within the AntiNex Core to make new predictions.
Train a New Deep Neural Network and Make Predictions¶
This uses environment variables to build, train and make predictions using the AntiNex client.
This python script is available in the pip: ai_train_dnn.py
It takes parameters:
parser = argparse.ArgumentParser(
description=("Python client to Train a Deep Neural Network "
"with AntiNex Django Rest Framework"))
parser.add_argument(
"-u",
help="username",
required=False,
dest="user")
parser.add_argument(
"-p",
help="user password",
required=False,
dest="password")
parser.add_argument(
"-e",
help="user email",
required=False,
dest="email")
parser.add_argument(
"-a",
help="url endpoint with default http://localhost:8010",
required=False,
dest="url")
parser.add_argument(
"-f",
help="file to use default ./examples/test-keras-dnn.json",
required=False,
dest="datafile")
parser.add_argument(
"-s",
help="silent",
required=False,
dest="silent",
action="store_true")
parser.add_argument(
"-d",
help="debug",
required=False,
dest="debug",
action="store_true")
Get an ML Job¶
This python script is available in the pip: ai_get_job.py
It takes parameters:
parser = argparse.ArgumentParser(
description=("Python client get AI Job by ID"))
parser.add_argument(
"-u",
help="username",
required=False,
dest="user")
parser.add_argument(
"-p",
help="user password",
required=False,
dest="password")
parser.add_argument(
"-e",
help="user email",
required=False,
dest="email")
parser.add_argument(
"-a",
help="url endpoint with default http://localhost:8010",
required=False,
dest="url")
parser.add_argument(
"-i",
help="User's MLJob.id to look up",
required=False,
dest="job_id")
parser.add_argument(
"-s",
help="silent",
required=False,
dest="silent",
action="store_true")
parser.add_argument(
"-d",
help="debug",
required=False,
dest="debug",
action="store_true")
Get ML Job Results¶
This python script is available in the pip: ai_get_results.py
It takes parameters:
parser = argparse.ArgumentParser(
description=("Python client get AI Results by ID"))
parser.add_argument(
"-u",
help="username",
required=False,
dest="user")
parser.add_argument(
"-p",
help="user password",
required=False,
dest="password")
parser.add_argument(
"-e",
help="user email",
required=False,
dest="email")
parser.add_argument(
"-a",
help="url endpoint with default http://localhost:8010",
required=False,
dest="url")
parser.add_argument(
"-i",
help="User's MLJobResult.id to look up",
required=False,
dest="result_id")
parser.add_argument(
"-s",
help="silent",
required=False,
dest="silent",
action="store_true")
parser.add_argument(
"-d",
help="debug",
required=False,
dest="debug",
action="store_true")
Prepare a New Dataset¶
This python script is available in the pip: ai_prepare_dataset.py
It takes parameters:
parser = argparse.ArgumentParser(
description=("Python client to Prepare a dataset"))
parser.add_argument(
"-u",
help="username",
required=False,
dest="user")
parser.add_argument(
"-p",
help="user password",
required=False,
dest="password")
parser.add_argument(
"-e",
help="user email",
required=False,
dest="email")
parser.add_argument(
"-a",
help="url endpoint with default http://localhost:8010",
required=False,
dest="url")
parser.add_argument(
"-f",
help="file to use default ./examples/test-keras-dnn.json",
required=False,
dest="prepare_file")
parser.add_argument(
"-s",
help="silent",
required=False,
dest="silent",
action="store_true")
parser.add_argument(
"-d",
help="debug",
required=False,
dest="debug",
action="store_true")
Get a Prepared Dataset from the Database¶
This python script is available in the pip: ai_get_prepared_dataset.py
It takes parameters:
parser.add_argument(
"-u",
help="username",
required=False,
dest="user")
parser.add_argument(
"-p",
help="user password",
required=False,
dest="password")
parser.add_argument(
"-e",
help="user email",
required=False,
dest="email")
parser.add_argument(
"-a",
help="url endpoint with default http://localhost:8010",
required=False,
dest="url")
parser.add_argument(
"-i",
help="User's MLPrepare.id to look up",
required=False,
dest="prepare_id")
parser.add_argument(
"-s",
help="silent",
required=False,
dest="silent",
action="store_true")
parser.add_argument(
"-d",
help="debug",
required=False,
dest="debug",
action="store_true")