From 5bf4ab50e13a9be2923ddebe69525b40df8ab3bc Mon Sep 17 00:00:00 2001
From: Sacha Chua <sacha@sachachua.com>
Date: Sat, 5 Nov 2022 08:02:44 -0400
Subject: Separate caption and prerec roles

---
 common-playbook.yml                           |  5 ++++
 group_vars/all.yml                            |  1 +
 roles/caption/defaults/main.yml               |  1 -
 roles/caption/tasks/main.yml                  | 14 ---------
 roles/caption/templates/process-prerec.sh     | 18 -----------
 roles/caption/templates/reencode.sh           | 43 ---------------------------
 roles/caption/templates/run-aeneas.sh         | 14 ---------
 roles/caption/templates/update-task-status.sh | 10 -------
 roles/caption/templates/upload.sh             |  6 ----
 roles/prerec/tasks/main.yml                   | 28 +++++++++++++++++
 roles/prerec/templates/process-prerec.sh      | 18 +++++++++++
 roles/prerec/templates/run-aeneas.sh          | 14 +++++++++
 roles/prerec/templates/update-task-status.sh  | 10 +++++++
 roles/prerec/templates/upload.sh              |  6 ++++
 14 files changed, 82 insertions(+), 106 deletions(-)
 delete mode 100755 roles/caption/templates/process-prerec.sh
 delete mode 100755 roles/caption/templates/reencode.sh
 delete mode 100755 roles/caption/templates/run-aeneas.sh
 delete mode 100755 roles/caption/templates/update-task-status.sh
 delete mode 100755 roles/caption/templates/upload.sh
 create mode 100644 roles/prerec/tasks/main.yml
 create mode 100755 roles/prerec/templates/process-prerec.sh
 create mode 100755 roles/prerec/templates/run-aeneas.sh
 create mode 100755 roles/prerec/templates/update-task-status.sh
 create mode 100755 roles/prerec/templates/upload.sh

diff --git a/common-playbook.yml b/common-playbook.yml
index a3377ea..463c411 100644
--- a/common-playbook.yml
+++ b/common-playbook.yml
@@ -41,6 +41,11 @@
         name: stream
         tasks_from: test
       when: icecast_test is defined and icecast_test=='file'
+- name: Set up prerec processing workflow
+  hosts: res
+  tags: prerec
+  roles:
+    - prerec
 - name: Set up captioning tools
   hosts: res
   tags: caption
diff --git a/group_vars/all.yml b/group_vars/all.yml
index 925a944..d3c2e7b 100644
--- a/group_vars/all.yml
+++ b/group_vars/all.yml
@@ -29,3 +29,4 @@ emacs_config_dir: ~{{ emacsconf_user }}/.emacs.d
 emacsconf_el_dir: ~{{ emacsconf_user }}/emacsconf-el
 emacsconf_edit_wiki_dir: ~{{ emacsconf_user }}/emacsconf-wiki
 emacsconf_private_dir: ~{{ emacsconf_user }}/emacsconf-2022-private
+emacsconf_caption_dir: /data/emacsconf/{{ emacsconf_year }}
diff --git a/roles/caption/defaults/main.yml b/roles/caption/defaults/main.yml
index c158118..8b13789 100644
--- a/roles/caption/defaults/main.yml
+++ b/roles/caption/defaults/main.yml
@@ -1,2 +1 @@
-emacsconf_caption_dir: /data/emacsconf/{{ emacsconf_year }}
 
diff --git a/roles/caption/tasks/main.yml b/roles/caption/tasks/main.yml
index a69d848..ab49c00 100644
--- a/roles/caption/tasks/main.yml
+++ b/roles/caption/tasks/main.yml
@@ -34,17 +34,6 @@
     state: directory
     owner: "{{ emacsconf_user }}"
     group: "{{ emacsconf_group }}"
-- name: Recreate encoding script
-  tags: process-prerec
-  template:
-    src: "{{ item }}"
-    dest: "{{ emacsconf_caption_dir }}/scripts/{{ item }}"
-    owner: "{{ emacsconf_user }}"
-    group: "{{ emacsconf_group }}"
-    force: no
-    mode: 0775
-  loop:
-    - reencode.sh
 - name: Copy scripts for processing
   tags: process-prerec
   template:
@@ -55,9 +44,6 @@
     mode: 0775
   loop:
     - process-captions.py
-    - process-prerec.sh
-    - update-task-status.sh
-    - upload.sh
     - run-aeneas.sh
 # - name: Copy the inotify script
 #   tags: process-captions
diff --git a/roles/caption/templates/process-prerec.sh b/roles/caption/templates/process-prerec.sh
deleted file mode 100755
index e49aa72..0000000
--- a/roles/caption/templates/process-prerec.sh
+++ /dev/null
@@ -1,18 +0,0 @@
-#!/bin/bash
-# {{ ansible_managed }}
-
-ORIGINAL=$1
-REENCODED=$(echo "$ORIGINAL" | perl -pe 's/^(emacsconf-[0-9]*-.*?--.*?--.*?--).*/$1reencoded.webm/')
-SLUG=$(echo "$ORIGINAL" | perl -ne '/^emacsconf-[0-9]*-(.*?)--/ && print $1')
-MAIN=$(echo "$ORIGINAL" | perl -pe 's/^(emacsconf-[0-9]*-.*?--.*?--.*?--).*/$1main.webm/')
-SCREEN=reencode-$SLUG
-if ! ( screen -ls | grep -q $SLUG ); then
-    screen -dmS $SCREEN
-fi
-( cd /data/emacsconf/cache; ./update-cache )
-/data/emacsconf/2022/scripts/update-task-status.sh $SLUG "WAITING_FOR_PREREC" "PROCESSING"
-#if [[ ! -f "$REENCODED" ]]; then
-screen -S $SCREEN -X screen -t reencode-$SLUG /bin/bash -c "/data/emacsconf/2022/scripts/reencode.sh \"$ORIGINAL\" \"$REENCODED\" && /data/emacsconf/2022/scripts/upload.sh $REENCODED && exec /bin/bash" &
-#fi
-screen -S $SCREEN -X screen -t captions-$SLUG /bin/bash -c "/data/emacsconf/2022/scripts/process-captions.py $(dirname $ORIGINAL); /data/emacsconf/2022/scripts/update-task-status.sh $SLUG PROCESSING TO_ASSIGN; exec /bin/bash"
-screen -x $SCREEN
diff --git a/roles/caption/templates/reencode.sh b/roles/caption/templates/reencode.sh
deleted file mode 100755
index e3a82eb..0000000
--- a/roles/caption/templates/reencode.sh
+++ /dev/null
@@ -1,43 +0,0 @@
-#!/usr/bin/env bash
-
-set -euo pipefail
-
-# Defaults
-q=32
-cpu=4
-time_limit=""
-print_only=false
-
-while getopts :q:c:t:s OPT; do
-    case $OPT in
-        q|+q)
-            q="$OPTARG"
-            ;;
-        c|+c)
-            cpu="$OPTARG"
-            ;;
-        t|+t)
-            time_limit="-to $OPTARG"
-            ;;
-        s)
-            print_only=true
-            ;;
-        *)
-            echo "usage: `basename $0` [+-q ARG] [+-c ARG} [--] ARGS..."
-            exit 2
-    esac
-done
-shift `expr $OPTIND - 1`
-OPTIND=1
-
-command="$(cat<<EOF
-ffmpeg -y -i "$1" $time_limit -c:v libvpx-vp9 -b:v 0 -crf $q -an -row-mt 1 -tile-columns 2 -tile-rows 2 -cpu-used $cpu -g 240 -pass 1 -f webm -threads $cpu /dev/null &&
-    ffmpeg -y -i "$1" $time_limit -c:v libvpx-vp9 -b:v 0 -crf $q -c:a libopus -row-mt 1 -tile-columns 2 -tile-rows 2 -cpu-used $cpu -pass 2 -g 240 -threads $cpu "$2"
-EOF
-)"
-
-if [ $print_only == true ]; then
-    echo "$command"
-else
-    eval "$command"
-fi
diff --git a/roles/caption/templates/run-aeneas.sh b/roles/caption/templates/run-aeneas.sh
deleted file mode 100755
index 6f40134..0000000
--- a/roles/caption/templates/run-aeneas.sh
+++ /dev/null
@@ -1,14 +0,0 @@
-#!/usr/bin/bash
-#
-# {{ ansible_managed }}
-#
-AUDIO=$(ls *.opus *.ogg | head -n1)
-BASE=$(echo $AUDIO | perl -pe 's/^(emacsconf-2022-.*?--.*?--.*?)--.*/$1/')
-echo $AUDIO
-echo $BASE
-if [ ! -f $BASE--whisper.vtt ]; then
-  cp ${BASE}--main.vtt ${BASE}--whisper.vtt
-fi
-python3 -m aeneas.tools.execute_task $AUDIO *.txt "task_language=eng|os_task_file_format=vtt|is_text_type=plain" ${BASE}--aeneas.vtt
-cp ${BASE}--aeneas.vtt ${BASE}--main.vtt
-/data/emacsconf/2022/scripts/upload.sh ${BASE}--main.vtt
diff --git a/roles/caption/templates/update-task-status.sh b/roles/caption/templates/update-task-status.sh
deleted file mode 100755
index 18b7c1b..0000000
--- a/roles/caption/templates/update-task-status.sh
+++ /dev/null
@@ -1,10 +0,0 @@
-#!/bin/bash
-# {{ ansible_managed }}
-SLUG="$1"
-FROM_STATUS="$2"
-TO_STATUS="$3"
-cd ~/emacsconf-2022-private
-git pull
-emacsclient --eval "(progn (emacsconf-update-talk-status \"$SLUG\" \"$FROM_STATUS\" \"$TO_STATUS\") (emacsconf-with-talk-heading \"$SLUG\" (emacsconf-cache-video-data (emacsconf-get-talk-info-for-subtree)) (save-buffer)))"
-git commit -m "Update task status for $SLUG" conf.org
-git push
diff --git a/roles/caption/templates/upload.sh b/roles/caption/templates/upload.sh
deleted file mode 100755
index f723d29..0000000
--- a/roles/caption/templates/upload.sh
+++ /dev/null
@@ -1,6 +0,0 @@
-#!/bin/bash
-# {{ ansible_managed }}
-
-scp $* orga@media.emacsconf.org:~/backstage
-emacsclient --eval "(emacsconf-publish-backstage-index)"
-rsync -avze ssh orga@media.emacsconf.org:~/backstage/ /data/emacsconf/cache/
diff --git a/roles/prerec/tasks/main.yml b/roles/prerec/tasks/main.yml
new file mode 100644
index 0000000..5241531
--- /dev/null
+++ b/roles/prerec/tasks/main.yml
@@ -0,0 +1,28 @@
+- name: Ensure the directory exists
+  file:
+    path: "{{ emacsconf_caption_dir }}/scripts"
+    state: directory
+    owner: "{{ emacsconf_user }}"
+    group: "{{ emacsconf_group }}"
+- name: Recreate encoding script only if it's not there
+  tags: process-prerec
+  template:
+    src: "reencode.sh"
+    dest: "{{ emacsconf_caption_dir }}/scripts/reencode.sh"
+    owner: "{{ emacsconf_user }}"
+    group: "{{ emacsconf_group }}"
+    # This avoids overwriting zaeph's
+    force: no
+    mode: 0775
+- name: Copy scripts for processing
+  tags: process-prerec
+  template:
+    src: "{{ item }}"
+    dest: "{{ emacsconf_caption_dir }}/scripts/{{ item }}"
+    owner: "{{ emacsconf_user }}"
+    group: "{{ emacsconf_group }}"
+    mode: 0775
+  loop:
+    - process-prerec.sh
+    - update-task-status.sh
+    - upload.sh
diff --git a/roles/prerec/templates/process-prerec.sh b/roles/prerec/templates/process-prerec.sh
new file mode 100755
index 0000000..e49aa72
--- /dev/null
+++ b/roles/prerec/templates/process-prerec.sh
@@ -0,0 +1,18 @@
+#!/bin/bash
+# {{ ansible_managed }}
+
+ORIGINAL=$1
+REENCODED=$(echo "$ORIGINAL" | perl -pe 's/^(emacsconf-[0-9]*-.*?--.*?--.*?--).*/$1reencoded.webm/')
+SLUG=$(echo "$ORIGINAL" | perl -ne '/^emacsconf-[0-9]*-(.*?)--/ && print $1')
+MAIN=$(echo "$ORIGINAL" | perl -pe 's/^(emacsconf-[0-9]*-.*?--.*?--.*?--).*/$1main.webm/')
+SCREEN=reencode-$SLUG
+if ! ( screen -ls | grep -q $SLUG ); then
+    screen -dmS $SCREEN
+fi
+( cd /data/emacsconf/cache; ./update-cache )
+/data/emacsconf/2022/scripts/update-task-status.sh $SLUG "WAITING_FOR_PREREC" "PROCESSING"
+#if [[ ! -f "$REENCODED" ]]; then
+screen -S $SCREEN -X screen -t reencode-$SLUG /bin/bash -c "/data/emacsconf/2022/scripts/reencode.sh \"$ORIGINAL\" \"$REENCODED\" && /data/emacsconf/2022/scripts/upload.sh $REENCODED && exec /bin/bash" &
+#fi
+screen -S $SCREEN -X screen -t captions-$SLUG /bin/bash -c "/data/emacsconf/2022/scripts/process-captions.py $(dirname $ORIGINAL); /data/emacsconf/2022/scripts/update-task-status.sh $SLUG PROCESSING TO_ASSIGN; exec /bin/bash"
+screen -x $SCREEN
diff --git a/roles/prerec/templates/run-aeneas.sh b/roles/prerec/templates/run-aeneas.sh
new file mode 100755
index 0000000..6f40134
--- /dev/null
+++ b/roles/prerec/templates/run-aeneas.sh
@@ -0,0 +1,14 @@
+#!/usr/bin/bash
+#
+# {{ ansible_managed }}
+#
+AUDIO=$(ls *.opus *.ogg | head -n1)
+BASE=$(echo $AUDIO | perl -pe 's/^(emacsconf-2022-.*?--.*?--.*?)--.*/$1/')
+echo $AUDIO
+echo $BASE
+if [ ! -f $BASE--whisper.vtt ]; then
+  cp ${BASE}--main.vtt ${BASE}--whisper.vtt
+fi
+python3 -m aeneas.tools.execute_task $AUDIO *.txt "task_language=eng|os_task_file_format=vtt|is_text_type=plain" ${BASE}--aeneas.vtt
+cp ${BASE}--aeneas.vtt ${BASE}--main.vtt
+/data/emacsconf/2022/scripts/upload.sh ${BASE}--main.vtt
diff --git a/roles/prerec/templates/update-task-status.sh b/roles/prerec/templates/update-task-status.sh
new file mode 100755
index 0000000..18b7c1b
--- /dev/null
+++ b/roles/prerec/templates/update-task-status.sh
@@ -0,0 +1,10 @@
+#!/bin/bash
+# {{ ansible_managed }}
+SLUG="$1"
+FROM_STATUS="$2"
+TO_STATUS="$3"
+cd ~/emacsconf-2022-private
+git pull
+emacsclient --eval "(progn (emacsconf-update-talk-status \"$SLUG\" \"$FROM_STATUS\" \"$TO_STATUS\") (emacsconf-with-talk-heading \"$SLUG\" (emacsconf-cache-video-data (emacsconf-get-talk-info-for-subtree)) (save-buffer)))"
+git commit -m "Update task status for $SLUG" conf.org
+git push
diff --git a/roles/prerec/templates/upload.sh b/roles/prerec/templates/upload.sh
new file mode 100755
index 0000000..f723d29
--- /dev/null
+++ b/roles/prerec/templates/upload.sh
@@ -0,0 +1,6 @@
+#!/bin/bash
+# {{ ansible_managed }}
+
+scp $* orga@media.emacsconf.org:~/backstage
+emacsclient --eval "(emacsconf-publish-backstage-index)"
+rsync -avze ssh orga@media.emacsconf.org:~/backstage/ /data/emacsconf/cache/
-- 
cgit v1.2.3