Serialize whole quote

[wolnelektury.git] / Dockerfile

diff --git a/Dockerfile b/Dockerfile
index f916be0..2fc76de 100644 (file)
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,24 +1,14 @@
-FROM python:3.8 AS base
+# Base system and requirements.
+
+FROM python:3.9-trixie AS base
+
+ARG UID=8100
+ARG GID=8100
 
 RUN     apt-get update && apt-get install -y \
        git \
        calibre \
        texlive-xetex texlive-lang-polish \
-       libespeak-dev
-
-COPY requirements/requirements.txt requirements.txt
-
-# numpy -> aeneas
-RUN pip install numpy
-RUN pip install aeneas
-
-RUN pip install --no-cache-dir -r requirements.txt
-RUN pip install --no-cache-dir \
-    psycopg2-binary \
-    django-debug-toolbar==3.2.2 \
-    python-bidi
-
-RUN     apt-get install -y \
        texlive-extra-utils \
        texlive-lang-greek \
        texlive-lang-other \
@@ -28,20 +18,111 @@ RUN     apt-get install -y \
        fonts-noto-core fonts-noto-extra
 
 
+COPY requirements/requirements.txt requirements.txt
+
+RUN pip install --no-cache-dir -r requirements.txt
+RUN pip install --no-cache-dir \
+    psycopg2-binary \
+    django-debug-toolbar==3.2.2
+
+RUN addgroup --gid $GID app && \
+    adduser --gid $GID --home /app --uid $UID app
+
+
 # fonts
-RUN cp -a /usr/local/lib/python*/site-packages/librarian/fonts /usr/local/share/fonts
+RUN cp -a /usr/local/lib/python*/site-packages/librarian/fonts /usr/share/fonts
 RUN fc-cache
 
+USER app
+
 WORKDIR /app/src
 
+RUN mkdir /app/.ipython
+
+
+
+# dev
+# Assumes code will be mounted.
 
 FROM base AS dev
 
 #RUN pip install --no-cache-dir coverage
 
 
-FROM base AS prod
+# prod-base
+# Installs code and production requirements.
+
+FROM base AS prod-base
+
+USER root
 
 RUN pip install --no-cache-dir gunicorn
 
+USER app
+
 COPY src /app/src
+
+
+# static-builder
+# Static files are collected here. Statics are then copied into an Nginx image,
+# and staticfiles.json manifest file is copied back into production image.
+
+FROM prod-base AS static-builder
+
+RUN python manage.py collectstatic --noinput
+
+
+# nginx
+# Contains static files and nginx config.
+
+FROM nginx:alpine AS nginx
+
+ARG UID=8100
+ARG GID=8100
+
+# COPY nginx.conf conf.d/...
+
+RUN addgroup -g $GID app && addgroup nginx app
+COPY --from=static-builder /app/var/static /app/var/static
+
+
+
+
+# Postgres with Polish hunspell dictionary withoud diacritics
+# to allow full text search without diacritics.
+
+# --- Stage 1: prepare dictionary files ---
+FROM python:3.13-slim-trixie AS polish-dict
+
+RUN apt-get update \
+ && apt-get install -y --no-install-recommends hunspell-pl \
+ && rm -rf /var/lib/apt/lists/*
+
+COPY scripts/fold_hunspell.py /usr/local/bin/fold_hunspell.py
+
+RUN set -eu; \
+    mkdir /out; cd /tmp; \
+    enc="$(sed -n 's/^SET[[:space:]]*//p' /usr/share/hunspell/pl_PL.aff | tr -d '\r')"; \
+    iconv -f "${enc:-UTF-8}" -t UTF-8 /usr/share/hunspell/pl_PL.dic > pl.dict; \
+    iconv -f "${enc:-UTF-8}" -t UTF-8 /usr/share/hunspell/pl_PL.aff > pl.affix; \
+    python3 /usr/local/bin/fold_hunspell.py pl.dict pl.affix \
+        /out/polish.dict /out/polish.affix; \
+    dpkg-query -W -f='hunspell-pl ${Version}\n' hunspell-pl > /out/polish.version
+
+# --- Stage 2: actual Postgres image with the dictionary ---
+FROM postgres:18 AS postgres
+
+COPY --from=polish-dict /out/ /tmp/polish-dict/
+RUN mv /tmp/polish-dict/* "/usr/share/postgresql/$PG_MAJOR/tsearch_data/" \
+ && rmdir /tmp/polish-dict
+
+COPY scripts/postgres-polish.sql /docker-entrypoint-initdb.d/10-postgres-polish.sql
+
+
+
+# prod
+# That's the production image with staticfiles.json manifest file added.
+
+FROM prod-base AS prod
+
+COPY --from=static-builder /app/var/static/staticfiles.json /app/var/static/staticfiles.json