From e3b3299acd04cf282dfb91d777b85bf8409a7b06 Mon Sep 17 00:00:00 2001
From: Jean-Baptiste Holcroft <jean-baptiste@holcroft.fr>
Date: Sep 24 2023 19:19:48 +0000
Subject: make TM generation based on config file


these file processing takes a lot of time and disk space, let's make this optional

---

diff --git a/Dockerfile b/Dockerfile
index 6adc98b..cc2e937 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -17,6 +17,7 @@ RUN mkdir -p /src/results /srpms; chmod g+rwX /srpms; chmod -R 1777 /tmp
 WORKDIR /src
 
 COPY *.py *.sh /src/
+COPY configuration.json /src/configuration.json
 COPY dnf /src/dnf/
 COPY website /src/website/
 RUN chmod -R g+rwX /src/website
diff --git a/build_tm.py b/build_tm.py
index 4be4d93..1459d3c 100755
--- a/build_tm.py
+++ b/build_tm.py
@@ -52,6 +52,14 @@ def main():
     if os.path.exists(tm_folder) is False:
         os.makedirs(tm_folder)
 
+    log.info("Check if translation memories should be generated")
+    with open("configuration.json", "r") as read_file:
+        tm_for_versions = json.load(read_file)["tm_for_versions"]
+        if args.results not in tm_for_versions:
+            log.info(f"Configurations says to build tm only for {tm_for_versions}, nothing to do for {args.results}")
+            log.info("done")
+            return
+
     log.info("Find detected languages")
 
     languages = [
diff --git a/build_website.py b/build_website.py
index 5df4da4..69d05db 100755
--- a/build_website.py
+++ b/build_website.py
@@ -184,14 +184,18 @@ def generate_static_pages_langs(release: str, code: str, content: dict, destinat
     data["lang_code"] = code
     data["now"] = datetime.datetime.utcnow()
     data["files"] = defaultdict(dict)
-    data["files"]["compendium"]["url"] = f"/{release}/{code}.po.gz"
-    data["files"]["compendium"]["size"] = os.path.getsize(os.path.join(tm_folder, f"{code}.po.gz"))
-    data["files"]["terminology"]["url"] = f"/{release}/{code}.terminology.po.gz"
-    data["files"]["terminology"]["size"] = os.path.getsize(os.path.join(tm_folder, f"{code}.terminology.po.gz"))
-    data["files"]["tmx"]["url"] = f"/{release}/{code}.tmx.gz"
-    data["files"]["tmx"]["size"] = os.path.getsize(os.path.join(tm_folder, f"{code}.tmx.gz"))
-    data["files"]["csv"]["url"] = f"/{release}/{code}.csv"
-    data["files"]["csv"]["size"] = os.path.getsize(os.path.join(static_tm_folder, f"{code}.csv"))
+    if os.path.exists(os.path.join(tm_folder, f"{code}.po.gz")):
+        data["files"]["compendium"]["url"] = f"/{release}/{code}.po.gz"
+        data["files"]["compendium"]["size"] = os.path.getsize(os.path.join(tm_folder, f"{code}.po.gz"))
+    if os.path.exists(os.path.join(tm_folder, f"{code}.terminology.po.gz")):
+        data["files"]["terminology"]["url"] = f"/{release}/{code}.terminology.po.gz"
+        data["files"]["terminology"]["size"] = os.path.getsize(os.path.join(tm_folder, f"{code}.terminology.po.gz"))
+    if os.path.exists(os.path.join(tm_folder, f"{code}.tmx.gz")):
+        data["files"]["tmx"]["url"] = f"/{release}/{code}.tmx.gz"
+        data["files"]["tmx"]["size"] = os.path.getsize(os.path.join(tm_folder, f"{code}.tmx.gz"))
+    if os.path.exists(os.path.join(static_tm_folder, f"{code}.csv")):
+        data["files"]["csv"]["url"] = f"/{release}/{code}.csv"
+        data["files"]["csv"]["size"] = os.path.getsize(os.path.join(static_tm_folder, f"{code}.csv"))
     if len(territories) > 0:
         data["territories"] = territories
 
diff --git a/configuration.json b/configuration.json
new file mode 100644
index 0000000..28b561b
--- /dev/null
+++ b/configuration.json
@@ -0,0 +1,4 @@
+{
+  "tm_for_versions_comment": "only build tm for this fedora version",
+  "tm_for_versions": ["f38"]
+}
\ No newline at end of file
diff --git a/templates/language.md b/templates/language.md
index bd3042e..566ba55 100644
--- a/templates/language.md
+++ b/templates/language.md
@@ -28,11 +28,21 @@ Possible scripts are: {% for script in scripts -%}{{ script }} {%- endfor %}
 * Source words to translate in started packages: {{ totalsourcewordssum }}
 * Translated words: {{ translatedsourcewords }}
 
+{% if files %}
 Download:
+{% if "compendium" in files %}
 * [{{ lang_code }} compendium ({{ files["compendium"]["size"]|filesizeformat() }})]({{ "{{% resource url=" }}"{{ files["compendium"]["url"] }}" {{ "%}}" }}) (aggregation of all strings found in po files)
+{% endif %}
+{% if "terminology" in files %}
 * [{{ lang_code }} terminology ({{ files["terminology"]["size"]|filesizeformat() }})]({{ "{{% resource url=" }}"{{ files["terminology"]["url"] }}" {{ "%}}" }}) see [poterminology](https://docs.translatehouse.org/projects/translate-toolkit/en/latest/commands/poterminology.html)
+{% endif %}
+{% if "tmx" in files %}
 * [{{ lang_code }} translation memory ({{ files["tmx"]["size"]|filesizeformat() }})]({{ "{{% resource url=" }}"{{ files["tmx"]["url"] }}" {{ "%}}" }}) see [tmx](https://en.wikipedia.org/wiki/Translation_Memory_eXchange)
+{% endif %}
+{% if "csv" in files %}
 * [{{ lang_code }} generated stats ({{ files["csv"]["size"]|filesizeformat() }})]({{ "{{% resource url=" }}"{{ files["csv"]["url"] }}" {{ "%}}" }})
+{% endif %}
+{% endif %}
 
 Packages: