From 4fc8c29ea8f36f79da95d440a9aad72e3db0ee78 Mon Sep 17 00:00:00 2001
From: Rodolfo Carvalho <rhcarvalho@gmail.com>
Date: Fri, 2 Sep 2022 14:03:04 +0200
Subject: [PATCH] Use GitHub Actions to update repo on schedule

---
 .github/workflows/update.yml | 64 ++++++++++++++++++++++++++++++++++++
 README.md                    |  2 ++
 meta.awk                     |  8 +++++
 release_date.awk             |  4 +++
 4 files changed, 78 insertions(+)
 create mode 100644 .github/workflows/update.yml
 create mode 100644 meta.awk
 create mode 100644 release_date.awk

diff --git a/.github/workflows/update.yml b/.github/workflows/update.yml
new file mode 100644
index 0000000..9b385e2
--- /dev/null
+++ b/.github/workflows/update.yml
@@ -0,0 +1,64 @@
+name: "Update CC-CEDICT"
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - "main"
+    paths:
+      - ".github/workflows/update.yml"
+      - "cedict_1_0_ts_utf-8_mdbg.txt"
+      - "*_test.go"
+  schedule:
+    - cron: "0 9 * * *"
+permissions:
+  contents: "write"
+defaults:
+  run:
+    shell: "bash"
+concurrency:
+  group: "update-${{ github.ref }}"
+  cancel-in-progress: true
+jobs:
+  update:
+    name: "Update CC-CEDICT"
+    runs-on: "ubuntu-latest"
+    timeout-minutes: 5
+    steps:
+      - uses: "actions/checkout@v3"
+      - name: "Configure Git"
+        run: |
+          git config user.name "Rodolfo Carvalho"
+          git config user.email "88819+rhcarvalho@users.noreply.github.com"
+      - name: "Download"
+        run: |
+          URL='https://www.mdbg.net/chinese/export/cedict/cedict_1_0_ts_utf-8_mdbg.txt.gz'
+          curl -L ${URL} | gunzip > cedict_1_0_ts_utf-8_mdbg.txt
+      - name: "Check for changes"
+        id: "diff"
+        run: |
+          git add -- cedict_1_0_ts_utf-8_mdbg.txt
+          git diff --cached --stat
+          echo "::set-output name=changed::$(git diff --cached --quiet; echo $?)"
+      - name: "Update metadata"
+        if: "fromJSON(steps.diff.outputs.changed)"
+        run: |
+          awk -f meta -v RS='\r\n' cedict_1_0_ts_utf-8_mdbg.txt >> meta.csv
+          git add -- meta.csv
+      - name: "Commit"
+        if: "fromJSON(steps.diff.outputs.changed)"
+        run: |
+          RELEASE_DATE="$(awk -f release_date -v RS='\r\n' cedict_1_0_ts_utf-8_mdbg.txt)"
+          git commit -m "CC-CEDICT ${RELEASE_DATE}"
+      - uses: "actions/setup-go@v3"
+        if: "fromJSON(steps.diff.outputs.changed)"
+        with:
+          go-version-file: "go.mod"
+          # cache: true
+          # cache-dependency-path: "go.sum"
+      - name: "Test"
+        if: "fromJSON(steps.diff.outputs.changed)"
+        run: "go test -v -count=1"
+      - name: Push
+        if: "fromJSON(steps.diff.outputs.changed)"
+        run: |
+          git push
diff --git a/README.md b/README.md
index df8b731..6454991 100644
--- a/README.md
+++ b/README.md
@@ -25,6 +25,8 @@ func main() {
 }
 ```
 
+The dictionary is automatically updated daily (if there are changes upstream).
+
 <!--
 | CC-CEDICT Release Date | Number of Entries |
 |------------------------|-------------------|
diff --git a/meta.awk b/meta.awk
new file mode 100644
index 0000000..05c8e2a
--- /dev/null
+++ b/meta.awk
@@ -0,0 +1,8 @@
+match($0, "^^#! (date|time|entries)=(.+)$", m) {
+    a[m[1]] = m[2]
+}
+
+/^[^#]/ {
+    print a["date"]","a["time"]","a["entries"]
+    exit
+}
diff --git a/release_date.awk b/release_date.awk
new file mode 100644
index 0000000..fab2a63
--- /dev/null
+++ b/release_date.awk
@@ -0,0 +1,4 @@
+match($0, "^^#! date=(.+)$", m) {
+    print m[1]
+    exit
+}