From dc9b743afac253f858eedbc944cfc944cc9ac959 Mon Sep 17 00:00:00 2001
From: Ben Knoll <benknoll@umn.edu>
Date: Mon, 3 Oct 2022 11:17:18 -0500
Subject: [PATCH] Downloads stanza models earlier

Also includes updated dockerfiles
---
 Dockerfile                                    | 19 +++++++++----
 Dockerfile-dev                                | 28 +++++++++++++++++++
 python/biomedicus/cli.py                      |  1 +
 .../deployment/biomedicus_deploy_config.yml   |  1 +
 .../deployment/deploy_biomedicus.py           | 12 +++++---
 tools/docker/biomedicus.sh                    |  6 ++++
 6 files changed, 57 insertions(+), 10 deletions(-)
 create mode 100644 Dockerfile-dev
 create mode 100644 tools/docker/biomedicus.sh

diff --git a/Dockerfile b/Dockerfile
index afe0df58..03f7b634 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,4 +1,4 @@
-FROM ubuntu:latest
+FROM ubuntu:20.04
 
 ARG DEBIAN_FRONTEND=noninteractive
 
@@ -9,11 +9,18 @@ RUN apt-get -qy install openjdk-11-jre-headless
 RUN apt-get -qy install build-essential
 RUN apt-get -qy install python3 python3-pip
 
-COPY . /home/biomedicus3/biomedicus3
-
 WORKDIR /home/biomedicus3
 
-RUN pip3 install ./biomedicus3
-RUN biomedicus download-data
+RUN pip3 install --upgrade pip
+RUN python3 -m pip install --upgrade setuptools
+
+RUN pip3 install biomedicus3
+RUN biomedicus download-data --with-stanza
+
+COPY ./python/biomedicus/deployment/biomedicus_deploy_config.yml .
+COPY ./python/biomedicus/pipeline/biomedicus_default_pipeline.yml .
+COPY ./tools/docker/biomedicus.sh .
+
+RUN chmod +x biomedicus.sh
 
-CMD /bin/sh
+ENTRYPOINT ./biomedicus.sh
diff --git a/Dockerfile-dev b/Dockerfile-dev
new file mode 100644
index 00000000..13e77c3a
--- /dev/null
+++ b/Dockerfile-dev
@@ -0,0 +1,28 @@
+FROM ubuntu:20.04
+
+ARG DEBIAN_FRONTEND=noninteractive
+
+RUN apt-get update
+
+RUN apt-get -qy install git
+RUN apt-get -qy install openjdk-11-jdk-headless
+RUN apt-get -qy install build-essential
+RUN apt-get -qy install python3 python3-pip
+
+WORKDIR /home/biomedicus3
+
+RUN pip3 install --upgrade pip
+RUN python3 -m pip install --upgrade setuptools
+
+COPY . biomedicus3
+
+RUN pip3 install ./biomedicus3
+RUN biomedicus download-data --with-stanza
+
+COPY ./python/biomedicus/deployment/biomedicus_deploy_config.yml .
+COPY ./python/biomedicus/pipeline/biomedicus_default_pipeline.yml .
+COPY ./tools/docker/biomedicus.sh .
+
+RUN chmod +x biomedicus.sh
+
+ENTRYPOINT ./biomedicus.sh
diff --git a/python/biomedicus/cli.py b/python/biomedicus/cli.py
index 8a3afc54..8fd2bcd7 100644
--- a/python/biomedicus/cli.py
+++ b/python/biomedicus/cli.py
@@ -83,6 +83,7 @@ def write_config_subparser(subparsers):
 def download_data_subparser(subparsers):
     from biomedicus.deployment.deploy_biomedicus import download_data
     sp = subparsers.add_parser('download-data', help="Just downloads the biomedicus data.")
+    sp.add_argument('--with-stanza', action='store_true', help="Also downloads stanza models.")
     sp.set_defaults(f=download_data)
 
 
diff --git a/python/biomedicus/deployment/biomedicus_deploy_config.yml b/python/biomedicus/deployment/biomedicus_deploy_config.yml
index c4fb83b3..8b1492fd 100644
--- a/python/biomedicus/deployment/biomedicus_deploy_config.yml
+++ b/python/biomedicus/deployment/biomedicus_deploy_config.yml
@@ -39,6 +39,7 @@ processors:
     entry_point: biomedicus.dependencies.stanza_selective_parser
     port: 50800
     startup_timeout: 180
+    additional_args: ['--offline']
   - implementation: python
     entry_point: biomedicus.negation.deepen
     port: 50900
diff --git a/python/biomedicus/deployment/deploy_biomedicus.py b/python/biomedicus/deployment/deploy_biomedicus.py
index 96a51fed..e32ccc44 100644
--- a/python/biomedicus/deployment/deploy_biomedicus.py
+++ b/python/biomedicus/deployment/deploy_biomedicus.py
@@ -38,13 +38,17 @@ def _listen(process: Popen) -> int:
     return process.wait()
 
 
-def check_data(download=False):
+def check_data(download=False, with_stanza=False):
     try:
         data = Path(os.environ['BIOMEDICUS_DATA'])
     except KeyError:
         data = Path.home() / '.biomedicus' / 'data'
         os.environ['BIOMEDICUS_DATA'] = str(data)
 
+    if with_stanza:
+        import stanza
+        stanza.download('en')
+
     config = load_config()
     download_url = config['data.data_url']
     data_version = config['data.version']
@@ -105,7 +109,7 @@ def attach_biomedicus_jar(deployment: Deployment, append_to: Optional[str] = Non
 
 def deploy(conf):
     try:
-        check_data(conf.download_data)
+        check_data(conf.download_data, with_stanza=True)
     except ValueError:
         return
     deployment = Deployment.from_yaml_file(conf.config)
@@ -118,9 +122,9 @@ def deploy(conf):
     deployment.run_servers()
 
 
-def download_data(_):
+def download_data(conf):
     try:
-        check_data(True)
+        check_data(True, conf.with_stanza)
     except ValueError:
         return
 
diff --git a/tools/docker/biomedicus.sh b/tools/docker/biomedicus.sh
new file mode 100644
index 00000000..63f09ae6
--- /dev/null
+++ b/tools/docker/biomedicus.sh
@@ -0,0 +1,6 @@
+#!/usr/bin/env bash
+
+exec 3< <(biomedicus deploy --config biomedicus_deploy_config.yml)
+sed '/Done deploying all servers.$/q' <&3 ; cat <&3 &
+
+biomedicus run /input --watch -o /output --config biomedicus_default_pipeline.yml