Build Docker image without ingestor dependencies

.dockerignore

@@ -5,6 +5,7 @@ aleph.env
 aleph.egg-info
 docker-compose.*
 celerybeat-schedul*
+contrib/allCountries.zip
 .git
 .vscode
 .travis.yml

.gitignore

@@ -16,6 +16,8 @@ ui/i18n/messages.json
 ui/i18n/messages.pot
 ui/**/*.css
 
+contrib/allCountries.zip
+
 celerybeat-schedule*
 aleph.env
 test_settings.py

Dockerfile

@@ -1,7 +1,22 @@
-FROM alephdata/aleph-base:10
+FROM ubuntu:19.04
+ENV DEBIAN_FRONTEND noninteractive
+
+RUN apt-get -qq -y update \
+    && apt-get -qq -y install build-essential locales \
+        ca-certificates postgresql-client \
+        python3-pip python3-dev python3-icu python3-psycopg2 \
+        python3-lxml python3-crypto python3-gunicorn \
+    && apt-get -qq -y autoremove \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
+
+# Set up the locale and make sure the system uses unicode for the file system.
+RUN localedef -i en_US -c -f UTF-8 -A /usr/share/locale/locale.alias en_US.UTF-8
+ENV LANG en_US.utf8
 
 # Install Python dependencies
-RUN pip3 install spacy==2.1.4
+RUN pip3 install --no-cache-dir -q -U pip setuptools six
+RUN pip3 install --no-cache-dir -q spacy==2.1.4
 RUN python3 -m spacy download xx
 COPY requirements-generic.txt /tmp/
 RUN pip3 install --no-cache-dir -r /tmp/requirements-generic.txt
@@ -16,9 +31,9 @@ RUN pip install -e /aleph
 # RUN cd /usr/local/lib/python3.6/dist-packages && python3 /aleph/setup.py develop
 
 # Configure some docker defaults:
-ENV C_FORCE_ROOT=true \
-    ALEPH_ELASTICSEARCH_URI=http://elasticsearch:9200/ \
+ENV ALEPH_ELASTICSEARCH_URI=http://elasticsearch:9200/ \
     ALEPH_DATABASE_URI=postgresql://aleph:aleph@postgres/aleph \
+    ALEPH_GEONAMES_DATA=/aleph/contrib/geonames.txt \
     REDIS_URL=redis://redis:6379/0 \
     ARCHIVE_PATH=/data
 

Makefile

@@ -67,6 +67,12 @@ fixtures:
 	aleph crawldir --wait -f fixtures aleph/tests/fixtures/samples
 	balkhash iterate -d fixtures >aleph/tests/fixtures/samples.ijson
 
+contrib/allCountries.zip:
+	curl -s -o contrib/allCountries.zip https://download.geonames.org/export/dump/allCountries.zip
+
+geonames: contrib/allCountries.zip
+	unzip -p contrib/allCountries.zip | grep "ADM1\|PCLI\|PCLD\|PPLC\|PPLA" >contrib/geonames.txt
+
 # pybabel init -i aleph/translations/messages.pot -d aleph/translations -l de -D aleph
 translate: dev
 	pybabel extract -F babel.cfg -k lazy_gettext -o aleph/translations/messages.pot aleph