From 6f78611ec0163caebd16c8852fc280ab9a6726cb Mon Sep 17 00:00:00 2001
From: ksqsf <i@ksqsf.moe>
Date: Thu, 30 Nov 2023 21:54:25 +0800
Subject: [PATCH] ci: build simplified moran

---
 .github/workflows/main.yml | 13 +++++++
 make_simp_dist.sh          | 78 ++++++++++++++++++++++++++++++++++++++
 moran_sentence.schema.yaml |  3 --
 tools/schemagen.py         |  1 -
 4 files changed, 91 insertions(+), 4 deletions(-)
 create mode 100644 .github/workflows/main.yml
 create mode 100755 make_simp_dist.sh

diff --git a/.github/workflows/main.yml b/.github/workflows/main.yml
new file mode 100644
index 0000000..16e85bf
--- /dev/null
+++ b/.github/workflows/main.yml
@@ -0,0 +1,13 @@
+on: [push]
+jobs:
+  simplified-moran:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - run: pip3 install opencc pypinyin regex
+      - run: sudo apt install opencc p7zip-full
+      - run: ./make_simp_dist.sh
+      - uses: actions/upload-artifact@v3
+        with:
+          name: "MoranSimplified"
+          path: MoranSimplified*.7z
diff --git a/make_simp_dist.sh b/make_simp_dist.sh
new file mode 100755
index 0000000..bc2e83e
--- /dev/null
+++ b/make_simp_dist.sh
@@ -0,0 +1,78 @@
+#!/bin/bash
+
+git archive HEAD -o archive.tar
+mkdir -p dist
+tar xf archive.tar -C dist
+rm archive.tar
+cd dist
+
+# 更新单字字频
+echo 更新单字字频...
+cd tools
+python3 schemagen.py --pinyin-table=./data/pinyin_simp.txt  update-char-weight --rime-dict=../moran.chars.dict.yaml > ../moran.chars.dict.yaml.bak
+mv ../moran.chars.dict.yaml{.bak,}
+cd ..
+
+
+# 替换辅助码
+echo 替换辅助码...
+compact_dicts=(
+    "moran.essay.dict.yaml"
+    "moran.tencent.dict.yaml"
+    "moran.moe.dict.yaml"
+    "moran.thuocl.dict.yaml"
+    "moran.computer.dict.yaml"
+    "moran.hanyu.dict.yaml"
+    "moran.words.dict.yaml"
+)
+
+simplifyDict() {
+    cp $1 $1.bak
+    opencc -c opencc/moran_t2s.json -i $1.bak -o $1
+    rm $1.bak
+}
+
+for dict in "${compact_dicts[@]}"; do
+    simplifyDict $dict
+done
+
+(cd tools/ && ./update_compact_dicts.sh)
+
+darwin=false;
+case "`uname`" in
+  Darwin*) darwin=true ;;
+esac
+
+sedi () {
+    case $(uname -s) in
+        *[Dd]arwin* | *BSD* ) sed -i '' "$@";;
+        *) sed -i "$@";;
+    esac
+}
+
+# 替換碼表
+echo 替換碼表...
+sedi 's|\&dict moran_fixed|\&dict moran_fixed_simp|' moran_fixed.defaults.yaml
+sedi 's|fixed/dictionary: moran_fixed|fixed/dictionary: moran_fixed_simp|' moran.defaults.yaml
+
+# 替换简体语法模型
+echo 替换简体语法模型...
+wget 'https://github.com/lotem/rime-octagram-data/raw/hans/zh-hans-t-essay-bgc.gram' -O zh-hans-t-essay-bgc.gram
+wget 'https://github.com/lotem/rime-octagram-data/raw/hans/zh-hans-t-essay-bgw.gram' -O zh-hans-t-essay-bgw.gram
+rm zh-hant-t-essay-bg{c,w}.gram
+for f in *.defaults.yaml 
+do
+    sedi 's/zh-hant-t-essay-bgw/zh-hans-t-essay-bgw/' $f
+    sedi 's/zh-hant-t-essay-bgc/zh-hans-t-essay-bgc/' $f
+done
+
+cd ..
+
+# 打包
+echo 打包...
+rm -rf dist/tools
+rm -rf dist/.git
+cp 简体版说明.txt 下载与安装说明.txt 更新纪要.txt dist
+
+7zz a -tzip -mx=9 -r "MoranSimplified-$(date +%Y%m%d).7z" dist
+rm -rf dist
diff --git a/moran_sentence.schema.yaml b/moran_sentence.schema.yaml
index b43ac7b..221047b 100644
--- a/moran_sentence.schema.yaml
+++ b/moran_sentence.schema.yaml
@@ -189,6 +189,3 @@ __patch:
   - moran.defaults.yaml:/patch?
   - moran_patches.yaml:/fix_opencc_t2s?
   - moran_sentence.custom.yaml:/patch?
-
-grammar:
-  language: zh-hant-t-essay-bgw
diff --git a/tools/schemagen.py b/tools/schemagen.py
index 4695af7..ff7f6f4 100644
--- a/tools/schemagen.py
+++ b/tools/schemagen.py
@@ -16,7 +16,6 @@
 from itertools import *
 import zrmify
 import flypyify
-import pandas
 import math
 import opencc
 from pypinyin import lazy_pinyin