bfergerson/gitsocratic

Source code query command line interface

GitSocratic

Source code query command line interface

GitSocratic provides a pain-free interface for querying source code. GitSocratic uses the omnilingual parser Babelfish to extract universal abstract syntax trees (which are additionally annotated with their semantic meanings) and imports that data into a knowledge graph powered by Grakn. This process is fully facilitated by GitSocratic's integration with Phenomena.

Features

• Supports question/answer source code queries
• Cross-language source code querying
• Semantic role querying
• UAST structure querying

Setup

GitSocratic requires access to two services:

• Babelfish
• Grakn

GitSocratic is able to install these services automatically using Docker. If you do no wish to use Docker you may supply the host and ports for these services through the config file or command.

To automatically setup these services with Docker, simply use the command:

gitsocratic init

Note: This command may take several minutes on the first use.

Bare-bones installation

apt-get update
apt install openjdk-8-jre-headless
apt install unzip
apt install docker.io
wget https://github.com/CodeBrig/GitSocratic/releases/download/v0.2.1-alpha/gitsocratic-0.2.1.zip
unzip gitsocratic-0.2.1.zip
cd gitsocratic-0.2.1/bin/
./gitsocratic init

Usage

gitsocratic [-hV] [-c=<configFile>] [COMMAND]

Commands

add-local-repo   Add local source code repository to the knowledge graph
add-remote-repo  Add remote source code repository to the knowledge graph
config           Configure GitSocratic
console          Open interactive source code query console
init             Initialize services necessary to use GitSocratic
logs             View logs for initialized services
query            Execute a single source code query
question         Execute a single source code question

More information: COMMANDS

License

Apache 2.0

gitsocratic's Issues

Use bblfshd image with drivers pre-installed

#11

Remove manually added Grakn binaries

Fill in all the todo: description

Add architecture design to readme

Record processing rate

Prettier error on no Docker

when not using both local babelfish and grakn

Travis-ci setup

Example queries

• functions which declare return value of Integer but return int

Automate release artifact uploads

Schema option

default

• full

options:

• necessary
• files
• functions

Support for Gists

Update readme queries to be questions

Project node

So files/methods can be traversed back to project

find/search command

Allows you to look through a given codebase/file(s) by Arthur filter logic

ConfigOption validator

Neo4j/Cypher Support

Ability to store just functions and their cyclomatic complexity

While making it flexible enough to create other custom filters

Upgrade to Phenomena v0.2

In readme show adding literals across languages

Dialogflow integration

Create dialog command and refactor question command to use Dialogflow instead

Upgrade to Grakn 1.5

Ability to generate visual chart/graph via query

Ability to reset config to default

Or just delete config file?

Release distribution including langauge drivers

Need to look into the feasibility of this

Ensure Java identifier access & method call observers work

Init Apache Skywalking & Source++ services

Ability to configure Grakn/Babelfish version

Embedded graph database mode

Provide ability to choose between stored and in-memory graph db

Most complex methods question

• what are the [limit] most complex methods?
• what are the [limit] most complex [language] methods?

Stats command

Show general statistics about stored source code:

• file count
• repo count
• method count
• languages
• SLOC
• project count
• etc

Upgrade to OmniSRC v0.3

Ability to ask questions on external projects

If question is asked on project not stored the project should be cloned and processed and then question answered. Further questions answered through locally stored data.

Upgrade to Grakn 1.5.7

Need to do this before releasing 0.2-alpha

Semantic role flags

add-remote-repo/add-local-repo

• individual
• actual

init

• individual
• actual

Query/question caching

remove-repo command

Example queries

Make equivalent example queries:

• http://codeontology.org/examples
• http://codeontology.org/sparql

Also, benchmark on JDK 8

ProGuard to make release distrubution smaller