Skip to content

The flask api for the pro-grepper. It preprocess the problem statement, fit the vectorizer and predict on the model (label powerset with random forest). It returns JSON object in response.

Notifications You must be signed in to change notification settings

pro-grepper-org/pro-grepper-api

Repository files navigation

Api

The model used for api is LabelPowerset Transformation with Random Forest Classifier

Route Information :

Api route : apiforprogrepper.herokuapp.com/api/companies

Rquest Type : POST

Params :

  • problem-statement: ' ' // problem statement text
  • tags : [ ] //list of tags default empty
  • num-companies: 5 //top n most probabilistic companies

Flow :

  1. Data received by the api is passed through the following functions :
  • HTML tags are removed if any
  • Punctuation symbols are removed if any
  • Other than alphabets, other symbols are removed
  • Stop words are removed
  • Stemming is done to convert the tokens into their base form
  1. The preprocessed data is then fit into the vectorizer (vectorizer_new.sav)

  2. Finally,the data is passed through the model(company_tags_lbps_lg_new.sav) and the probabilities are calculated.

  3. The top n probabilistic companies are returned.

About

The flask api for the pro-grepper. It preprocess the problem statement, fit the vectorizer and predict on the model (label powerset with random forest). It returns JSON object in response.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages