From 422a73ad7dc033b7e5e24a44a3b136351b86da1e Mon Sep 17 00:00:00 2001
From: "Aaron (\"AJ\") Steers" <aj@airbyte.io>
Date: Tue, 29 Apr 2025 18:11:25 -0700
Subject: [PATCH] chore: add poe `get-modified-connectors` task (#59130)

---
 poe-tasks/README.md                  |  5 ++++
 poe-tasks/get-modified-connectors.sh | 41 ++++++++++++++++++++++++++++
 poe_tasks.toml                       |  5 ++++
 3 files changed, 51 insertions(+)
 create mode 100644 poe-tasks/README.md
 create mode 100755 poe-tasks/get-modified-connectors.sh

diff --git a/poe-tasks/README.md b/poe-tasks/README.md
new file mode 100644
index 00000000000..681d07e9cc0
--- /dev/null
+++ b/poe-tasks/README.md
@@ -0,0 +1,5 @@
+# Poe Tasks Directory
+
+This directory stores scripts and task definitions that are intended to be executed from the PoeThePoet task manager.
+
+You may be able to run them directly, although it is probably easier to run via `poe`.
diff --git a/poe-tasks/get-modified-connectors.sh b/poe-tasks/get-modified-connectors.sh
new file mode 100755
index 00000000000..ffa459f9e42
--- /dev/null
+++ b/poe-tasks/get-modified-connectors.sh
@@ -0,0 +1,41 @@
+#!/usr/bin/env bash
+
+# This script is used to find all modified connector directories in the Airbyte repository.
+# It compares the current branch with the default branch and filters out files that match certain ignore patterns.
+
+DEFAULT_BRANCH="master"
+
+# 1) update remote default branch quietly
+git fetch --quiet origin $DEFAULT_BRANCH
+
+# 2) set up ignore patterns
+ignore_patterns=(
+  'README.md'
+  '.coveragerc'
+)
+# join with | into a grouped regex
+ignore_globs="($(IFS='|'; echo "${ignore_patterns[*]}"))$"
+# 3) collect all file changes
+committed=$(git diff --name-only origin/$DEFAULT_BRANCH...HEAD)
+staged=$(git diff --cached --name-only)
+unstaged=$(git diff --name-only)
+untracked=$(git ls-files --others --exclude-standard)
+
+# 4) merge into one list
+all_changes=$(printf '%s\n%s\n%s\n%s' "$committed" "$staged" "$unstaged" "$untracked")
+
+# 5) drop ignored files
+filtered=$(printf '%s\n' "$all_changes" | grep -v -E "/${ignore_globs}")
+
+# 6) keep only connector paths
+connectors=$(printf '%s\n' "$filtered" | grep -E '^airbyte-integrations/connectors/[^/]+' )
+
+# 7) extract just the connector directory name
+dirs=$(printf '%s\n' "$connectors" \
+  | sed -E 's|airbyte-integrations/connectors/([^/]+).*|\1|'
+)
+
+# 8) output sorted, unique list
+if [ "$dirs" ]; then
+  printf '%s\n' "$dirs" | sort -u
+fi
diff --git a/poe_tasks.toml b/poe_tasks.toml
index 3a739169345..1e2124697ec 100644
--- a/poe_tasks.toml
+++ b/poe_tasks.toml
@@ -22,6 +22,11 @@ pnpm install
 pnpm build
 '''
 
+[tasks.get-modified-connectors]
+help = """List Airbyte connectors modified. This considers all committed, staged, unstaged, and untracked files.
+"""
+shell = "./poe-tasks/get-modified-connectors.sh"
+
 [tasks.connector]
 help = "Run a poe task in a connector's directory. Usage: poe connector <connector-name> <task> [args...]"