django-google-prediction is a Django application that wraps the Google Prediction API to make it easier to build web applications with machine learning and data analysis functionality.
- Go to the Google Developers Console, click Create Project and fill in the project name and id.
- Go to your project, select APIs & auth > APIs and turn the Prediction API on.
- In your project dashboard, select Cloud Storage > Storage browser.
- You might be requested to enable billing for this feature.
- Click New Bucket, fill in as desired and upload your dataset.
- In your project dashboard, select APIs & auth > Credentials.
- Click Create New Client ID, select the Service account option and save your private key as private-key.p12 in your project's base directory.
- Still in that page, note the Email address provided for that key (e.g. 1234567890-abcdefghijklmnopqrstuvwxyz1234567890@developer.gserviceaccount.com).
- Add the following to settings.py:
import os
BASE_DIR = os.path.dirname(os.path.dirname(__file__))
GOOGLE_PREDICTION_PRIVATE_KEY = os.path.join(BASE_DIR, 'private-key.p12')
GOOGLE_PREDICTION_PROJECT_EMAIL = '_________' # REPLACE WITH YOUR PROJECT EMAIL- Add the google_prediction folder to your project's base directory.
- Install all the modules in requirements.txt (use pip install -r requirements.txt).
For a complete example, check django-google-prediction-example.
- Using Hosted Models
HostedModel('sample.languageid').predict('Me llamo Gabriel. Como te llamas?')
# {u'kind': u'prediction#output', u'outputLabel': u'Spanish', u'id': u'sample.languageid', u'selfLink': u'https://www.googleapis.com/prediction/v1.6/projects/414649711441/hostedmodels/sample.languageid/predict', u'outputMulti': [{u'score': u'0.032187', u'label': u'English'}, {u'score': u'0.512064', u'label': u'Spanish'}, {u'score': u'0.455749', u'label': u'French'}]}
m = HostedModel('sample.languageid')
m.predict('My name is Gabriel - nice to meet you!')['outputLabel']
# u'English'
m.predict("Je m'appelle Gabriel. J'aime la France!")['outputLabel']
# u'French'- Using Trained Models
# Listing all your models
TrainedModel.list("YOUR_PROJECT_ID")
# Creating a new model
# (DATASET_LOCATION must be in a Google Storage bucket; leave blank for empty model)
TrainedModel("YOUR_PROJECT_ID", "YOUR_MODEL_NAME").insert("DATASET_LOCATION")
# Inserting data into your model
TrainedModel("YOUR_PROJECT_ID", "YOUR_MODEL_NAME").update(OUTPUT, INPUT_DATA)
# Getting information about your model
TrainedModel("YOUR_PROJECT_ID", "YOUR_MODEL_NAME").get()
# Analyzing your model
TrainedModel("YOUR_PROJECT_ID", "YOUR_MODEL_NAME").analyze()
# Predicting based on your model
TrainedModel("YOUR_PROJECT_ID", "YOUR_MODEL_NAME").predict(INPUT_DATA)
# Deleting your model
TrainedModel("YOUR_PROJECT_ID", "YOUR_MODEL_NAME").delete()
# Multiple calls
m = TrainedModel("YOUR_PROJECT_ID", "YOUR_MODEL_NAME")
m.predict(INPUT_DATA)
m.predict(SOME_OTHER_DATA)
m.predict(MORE_STUFF_HERE)
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent