日本語版

This is a search system on Jupyter notebook.

Input a query through a browser interface and output Top-10 similar Jupyter notebook.

If the demo is not displayed, see retrieval_system/images/Demo_gif3_2.gif on this repository or following files on google drive.

demo 1

demo 2

Experimental evaluation

• DBMS: PostgreSQL, Neo4j, SQLite

• Jupyter Notebook

• This search system

Jupyter notebooks converted into the particular formats must be stored in Databases.

Clone this system by running following command:

git clone https://github.com/OnizukaLab/Similarity_Search_on_Computational_Notebooks.git

A file tree consisting of important files is as follows:

Similarity_Search_on_Computational_Notebooks/

├── retrieval_system/

│ ├── manage.py

│ ├── interface/

│ └── retrieval_system/

└── sample_dataset/

│ ├── neo4j_sample.zip

│ ├── data1.zip

│ ├── data2.zip

│ └── data3.zip

└── README.md

• Import postgres_sample.sql into postgres.

• Unzip sample_dataset/neo4j_sample.zip and move it into neo4j's data directory (e.g. /usr/local/var/neo4j/data).

• Make a directory notebooks_data, then unzip zipfiles in Similarity_Search_on_Computational_Notebooks/sample_dataset/ and put '.ipynb' files into notebooks_data/.

Navigate to notebooks_data/ and start followings:

• PostgreSQL

• Neo4J (localhost:7474)

• Jupyter Notebook (localhost:8888)

Navigate to Similarity_Search_on_Computational_Notebooks/retrieval_system/ and run following command to start our system.

python manage.py runserver <port>

Then use interface by accessing http://127.0.0.1:port/interface/ .

For example, if you want use port 8080, run following command:

python manage.py runserver 8080

and access http://127.0.0.1:8080/interface/ .