ResumedocxAnalyzer

A Web App connected to Cosmos DB to search and retrieve CV by keyword to help the screening process for HR. Tutorial of connection between Node.js app and Cosmos DB first.

The data folder

First, create a data folder (may be autogenerated), with subfolder input, output and schema. The input will contain the CVs that have been uploaded to the blob in Azure. The output folder will contain the entity partition of the CV (and raw text as well). The schema folder will also contain the schema of the different entities that are extracted in the CV. This allows to check that the desired entities were extracted in the different CVs.

Connection with Azure

This back-end for the web application (in my repository searchapp) uses 3 resources in Azure : Blob storage account, Cosmos DB and Azure Search. You can find documentation in the Azure documentation to create these different resources in the Azure portal.

In my repository I have removed the configuration files for the connection with these different resources. Therefore, in the bin folder, once the resources have been created in Azure, add 3 python files : config_blob.py, config_cosmos.py and config_search.py. Then copy and paste the different lines in the different files with your specific information.

In config_blob.py :

BLOB_NAME = "Add blob account name here"
BLOB_KEY = "Your storage account key 1"
BLOB_CONTAINER = "your container name"

In config_cosmos.py

CSRF_ENABLED = True

COSMOSDB_HOST = 'Link of the cosmos DB host'
COSMOSDB_KEY = 'write Cosmos DB key'

COSMOSDB_DATABASE = 'write the name of the Cosmos DB database'
COSMOSDB_COLLECTION = 'write the name of the Cosmos DB collection'

CONNECTION_STRING_SEARCH='AccountEndpoint='+COSMOSDB_HOST+';AccountKey='+COSMOSDB_KEY+';Database='+COSMOSDB_DATABASE

In config_search.py

SEARCH_NAME="your Azure search name"
API_VERSION="2017-11-11" #you can change it if you prefer another version. This is the latest version.
API_KEY="your Azure search API key"

DATASOURCE_NAME="your datasource name"
INDEX_NAME="your index name"
INDEXER_NAME="your indexer name"