From 989fd51f79ce3c0e5b7fc09840ea1ba3a78dd7f8 Mon Sep 17 00:00:00 2001
From: Vadim Nifadev <36514612+nifadyev@users.noreply.github.com>
Date: Fri, 4 Oct 2024 17:18:19 +0300
Subject: [PATCH] #78: Add pr diff size checker in CI and pre-commit (#91)

---
 .github/actions/large-pr-check/action.yaml | 49 ++++++++++++++++++++++
 .github/workflows/pr.yml                   | 16 +++++++
 .pre-commit-config.yaml                    |  7 ++++
 scripts/large-pr-checker.sh                | 30 +++++++++++++
 4 files changed, 102 insertions(+)
 create mode 100644 .github/actions/large-pr-check/action.yaml
 create mode 100644 scripts/large-pr-checker.sh

diff --git a/.github/actions/large-pr-check/action.yaml b/.github/actions/large-pr-check/action.yaml
new file mode 100644
index 0000000..e4414f9
--- /dev/null
+++ b/.github/actions/large-pr-check/action.yaml
@@ -0,0 +1,49 @@
+name: "Large PR checker"
+description: "Blocks PR if number of lines changed is excessive. Modified version of https://github.com/adolfosilva/gh-large-pr-check/blob/main/action.yml"
+
+inputs:
+  max_lines_changed:
+    description: "Maximum number of lines changed allowed"
+    required: true
+    default: "500"
+  target_branch:
+    description: The branch to compare against
+    required: true
+    default: main
+outputs:
+  total_lines_changed:
+    description: "Total lines changed in this PR"
+    value: ${{ steps.get_total_lines_changed.outputs.total_lines_changed }}
+
+runs:
+  using: "composite"
+  steps:
+    - id: fetch_target_branch
+      run: |
+        git fetch origin ${{ inputs.target_branch }}
+      shell: bash
+    - id: get_total_lines_changed
+      run: |
+        size=$(git diff --shortstat origin/${{ inputs.target_branch }} ':(exclude)*.lock' \
+        | awk '{ print $4+$6 }' \
+        | awk -F- '{print $NF}' \
+        | bc)
+
+        echo ""
+        echo "Total lines changed (note: *.lock files are excluded from this count): $size"
+        echo "total_lines_changed=$size" >> $GITHUB_OUTPUT
+      shell: bash
+    - name: Comment PR
+      if: ${{ fromJSON(steps.get_total_lines_changed.outputs.total_lines_changed) > fromJSON(inputs.max_lines_changed) }}
+      uses: thollander/actions-comment-pull-request@v2
+      with:
+        comment_tag: pr_size
+        mode: recreate
+        message: |
+          :boom: :boom: :boom:
+          Total lines changed ${{ steps.get_total_lines_changed.outputs.total_lines_changed }} is greater than ${{ inputs.max_lines_changed }}.
+          Please consider breaking this PR down.
+    - id: fail
+      if: ${{ fromJSON(steps.get_total_lines_changed.outputs.total_lines_changed) > fromJSON(inputs.max_lines_changed) }}
+      run: exit 1
+      shell: bash
diff --git a/.github/workflows/pr.yml b/.github/workflows/pr.yml
index 0eda5fd..77f3152 100644
--- a/.github/workflows/pr.yml
+++ b/.github/workflows/pr.yml
@@ -18,6 +18,19 @@ env:
   RUFF_VERSION: "0.6.7"
 
 jobs:
+  check-pr-diff:
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Check PR diff size
+        uses: ./.github/actions/large-pr-check
+        with:
+          target_branch: ${{ github.event.pull_request.base.ref }}
+          max_lines_changed: 300
+
   lint:
     runs-on: ubuntu-latest
     timeout-minutes: 10
@@ -37,6 +50,9 @@ jobs:
       - name: Run Ruff
         run: ruff check --output-format=github .
 
+      - name: Spell Check
+        uses: crate-ci/typos@master
+
   test:
     runs-on: ubuntu-latest
     strategy:
diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
index e187218..9c03f31 100644
--- a/.pre-commit-config.yaml
+++ b/.pre-commit-config.yaml
@@ -41,3 +41,10 @@ repos:
         stages: [pre-push]
         language: python
         types: [python]
+      - id: check-pr-size
+        name: check-pr-size
+        entry: chmod +x scripts/large-pr-checker.sh && ./large-pr-checker.sh
+        pass_filenames: false
+        stages: [pre-push]
+        language: system
+        types: [python]
diff --git a/scripts/large-pr-checker.sh b/scripts/large-pr-checker.sh
new file mode 100644
index 0000000..ef66b15
--- /dev/null
+++ b/scripts/large-pr-checker.sh
@@ -0,0 +1,30 @@
+#!/bin/bash
+
+# Script helps to block Pull Requests with exceeded number of changed lines of code.
+# Firstly, it compares HEAD with source branch
+# Then it counts changed lines of code, excluding specified files and respecting max lines of code
+
+# Read first input arg or take default value - 500
+MAX_CHANGED_LOC="${1:-500}"
+
+# Second argument (source branch) is required
+if [ -z "$2" ]; then
+  echo "No remote source branch supplied"
+  exit 1
+else
+  source_branch=$2
+fi
+
+# Parse such line `2 files changed, 18 insertions(+), 248 deletions(-)` and write to variable
+changed_loc=$(git diff --shortstat ${source_branch} ':(exclude)*.lock' \
+        | awk '{ print $4+$6 }' \
+        | awk -F- '{print $NF}' \
+        | bc)
+
+if [[ $changed_loc -le $MAX_CHANGED_LOC ]]; then
+  echo "✅ ${changed_loc} lines of code changed. It is allowed to create Pull Request"
+  exit 0
+else
+  echo "❌ ${changed_loc} lines of code changed. It is more than allowed ${MAX_CHANGED_LOC}. Please divide changes into several branches"
+  exit 1
+fi