1. git clone https://github.com/ian-jooble/jooble-ir-2018-rock.git
2. pip install requrements.txt
3. Run services mentioned above python "server_name".py
4. python client.py

Search service : main manager for all workflow.

Port: port 13565

Input paramets:

• query - query from user
• max_docs - maximum number of documents returned by the search engine

Actions:

1. import analyzer -> create prepared query (list of terms).
2. request to service ReverseIndex -> receive json with key "processed_data" - dictionary with docID's and texts.
3. request to service Snippets -> receive json with key "processed_data" - dictionary with docID's and snippets.

Return:

• documents - dictionary with docID's and snippets.

Service Indexer : executes search by term in inverted index.

Port: 13538

Input:

• data - list of terms
• max_docs - maximum number of documents returned by the search engine.

Actions:

1. You need to create inverted_index from documents, idf and normilized_tf_idf before starting to work with a search engine.
2. request to service Ranking -> receive json with key "ranked" - list with ranked docIDs.

Return: processed_data - dictionary with docID's and texts (number of docID == max_docs). {'id': docID, 'text': text}

Service Ranking : executes ranking by using cosine between two vectors.

Port: 13541

Before using this service for ranking you must create idf and normalized tf_idf from inverted index by using ReverseIndex Service.

Input:

• documents - list with relevant docID
• words - list with tokens from the query.

Return: ranked - list with ranked docIDs

Snippets service : creates snippet - first sentence which contains any term from query.

Port: 13542

Input:

• data - dictionary with docID's and texts (number of docID == max_docs).

Return: processed_data - dictionary with docID's and snippets. {'id': docID, 'snippet': snippet}

Service Client : creates UI for receiving user queries and displaying search results.

Port: 13560

You can change 'max_docs' here. Actions:

1. import SpellChecker -> checks query for mistakes and returns correct query if any were found.
2. request to service Search -> recieve json with key "documents" - dictionary with docID's and snippets.