Assisted Code Review with source{d} Lookout

Vadim Markovtsev

Machine Learning for

Large Scale Code Analysis

Plan

Origins
What is Lookout?
SDK
From scratch: typo correction

Origins

Many efforts target boring stuff

Boring means automatable

Automatable ≠ unattended 😔

When to help?

While you type = IDE
While you check = CI
While you review = PR
Periodically, asynchronously = cron

IDE

Instant feedback
Rich information
Rich UI
Instant feedback
Many different IDEs and languages

CI

Part of the workflow
More time to run
No UI
Must not be wrong
Longer feedback loop

PR reviews

Part of the workflow
More time to run
UI
Should not be wrong
Longer feedback loop

cron

Plenty of time to run
Outside of the workflow
No UI
Longest feedback loop

Comparison

	∈ workflow	UI	Feedback	Rich info	FP requirements
IDE	✓	✓✓	instant	✓	low
CI	✓	✗	minutes	✗	0
PR	✓	✓	minutes	✗	lowest
cron	✗	✗	hours	✗	low

When to help?

While you type = IDE
While you check = CI
While you review = PR
Periodically, asynchronously

What is Lookout?

Goals

Assisted code review platform
Tight git/GitHub integration
Analyzed language agnostic
Implementation language agnostic
Batteries included

Architecture

Push event

PR event

docs.sourced.tech/lookout

SDK

src-d/lookout-sdk

Single source of gRPC definitions
Low-level API: Go, Python
Low-level examples

src-d/lookout-sdk-ml

High-level Python API
Stateful analyzers
Integrated with source{d} ml ecosystem

Rule of 👍

High-level API

class MyAnalyzer(Analyzer):
    @classmethod
    def train(cls, ...) -> AnalyzerModel:
        # ...

    def analyze(self, ...) -> [Comment]:
        # do something with self.model

Train

@with_uasts_and_contents
def train(cls,
          ptr: ReferencePointer,
          config: Dict[str, Any],
          data_service: DataService,
          files: Iterable[File]
          ) -> AnalyzerModel:

Analyze

@with_changed_uasts_and_contents
def analyze(self,
            ptr_from: ReferencePointer,
            ptr_to: ReferencePointer,
            data_service: DataService,
            changes: Iterable[Change]
            ) -> [Comment]:

Behind the scenes

gRPC servers and clients
Pooling and threading
Database of trained models
Caches
Logging
Metrics

Hands on

Objective

class Packfile:

Objective

def read_packfile(...):
    # ...
 
class Packfile:

Objective

def read_packfile(...):
        # ...
     
    class Packfle:

Steps

Install lookout-sdk-ml and autocorrect
Write typos.py
Fork a repo, create a PR
Generate a new GitHub Personal Access Token
Run our analyzer

Summary

source{d} decided to build products for assisted code review
Lookout is a great platform for assisted code review
Writing new analyzers is easy and fun

Thank you

bit.ly/2H8U5ra