cli-searcher

You will find a lot of comments explaining the reasoning behind the implementation.

Run

With maven support, compile then run

mvn compile
mvn exec:java -Dexec.mainClass=searcher.Searcher -Dexec.args="filesystem"
mvn test -Dtest=searcher.TestSuite

Objective

We open up a tree pointing to the specified directory / path.

.cli-searcher
+-- filesystem
|   +-- a.txt
|   +-- b.json
|   +-- subdirectory
        |   +-- c.txt

Run the cli on /filesystem

java -jar Rank.jar Ranker /filesystem

We obtain a list of files:

files = [a.txt, b.json, c.txt]

We then want to run a query with a text expression.

And find similiarty among all this files based on it.

Ranker> some text to find and compare among files
a.txt:100%
b.json:90%
c.txt:0%
Ranker> new text to find and compare among files
a.txt:80%
b.json:70%
c.txt:65%
Ranker> :quit

Pipeline

1. We read the filesystem and obtain all documents
2. Read input from user, text_search to base similarity upon
3. Vectorize, apply word embedding for each document
4. Apply similarity algorithm