2016-AUG-21

config/services/containers/app/papra/papra.container.j2

@@ -14,6 +14,7 @@ PublishPort={{ services['papra']['ports']['http'] }}:1221/tcp
 
 # Volumes
 Volume=%h/data/containers/papra/data:/app/app-data:rw
+Volume=%h/containers/papra/db:/app/app-data/db:rw
 Volume=%h/data/containers/papra/backup:/app/backup:rw
 
 # General

config/services/containers/app/papra/script/papra-date-change.sh

@@ -0,0 +1,210 @@
+#!/bin/bash
+
+set -Eeuo pipefail
+
+URL=""
+ORG_ID=""
+TOKEN=""
+apply=false
+
+page_size=100
+page_index=0
+documents_read=0
+matched=0
+unmatched=0
+updated=0
+failed=0
+
+# usage() function
+usage() {
+    echo "Usage: $0 -u \"URL\" -o \"ORG_ID\" -t \"TOKEN\" [-a]"
+    echo "-u <URL>: Specify the URL of Papra, including https://"
+    echo "-o <ORG_ID>: Specify the organization ID of Papra"
+    echo "-t <TOKEN>: Specify the token to access the Papra API"
+    echo "-a: Apply changes. Without this option, only perform a dry run"
+    echo "Token requires documents:read and documents:update permissions"
+    exit 1
+}
+
+# Log function
+log() {
+    local timestamp
+    local level="$1"
+    local msg="$2"
+
+    timestamp="$(date -u +"%Y-%m-%dT%H:%M:%SZ")"
+    echo "time=\"$timestamp\" level=\"$level\" msg=\"$msg\" source=\"papra-date-change.sh\"" >&2
+}
+
+# getopts to get arguments
+while getopts ":u:o:t:ah" opt; do
+    case "$opt" in
+        u)
+            URL="$OPTARG"
+            ;;
+        o)
+            ORG_ID="$OPTARG"
+            ;;
+        t)
+            TOKEN="$OPTARG"
+            ;;
+        a)
+            apply=true
+            ;;
+        h)
+            usage
+            ;;
+        \?) # unknown option
+            log "error" "Invalid option: -$OPTARG"
+            usage
+            ;;
+        :) # parameter required
+            log "error" "Option -$OPTARG requires an argument"
+            usage
+            ;;
+    esac
+done
+
+# Validate required arguments
+if [[ -z "$URL" || -z "$ORG_ID" || -z "$TOKEN" ]]; then
+    log "error" "URL, ORG_ID and TOKEN are required"
+    usage
+fi
+
+# Require an explicit HTTPS URL
+if [[ "$URL" != https://* ]]; then
+    log "error" "URL must start with https://"
+    exit 1
+fi
+
+# Remove trailing slash
+URL="${URL%/}"
+
+# Collect all documents before modifying them.
+# This prevents pagination results from changing during updates.
+documents_file="$(mktemp)"
+trap 'rm -f -- "$documents_file"' EXIT
+
+while true; do
+    response="$(
+        curl \
+            --fail \
+            --silent \
+            --show-error \
+            --get \
+            --header "Authorization: Bearer $TOKEN" \
+            --data-urlencode "pageIndex=$page_index" \
+            --data-urlencode "pageSize=$page_size" \
+            --data-urlencode "sortField=createdAt" \
+            --data-urlencode "sortOrder=asc" \
+            "$URL/api/organizations/$ORG_ID/documents"
+    )"
+
+    count="$(jq -r '.documents | length' <<<"$response")"
+    total="$(jq -r '.documentsCount' <<<"$response")"
+
+    jq -c '.documents[]' <<<"$response" >>"$documents_file"
+
+    documents_read=$((documents_read + count))
+
+    if (( count == 0 || documents_read >= total )); then
+        break
+    fi
+
+    page_index=$((page_index + 1))
+done
+
+log "info" "Documents read: $documents_read"
+
+while IFS= read -r document; do
+    id="$(jq -r '.id' <<<"$document")"
+    name="$(jq -r '.name // empty' <<<"$document")"
+
+    if [[ -z "$id" || -z "$name" ]]; then
+        unmatched=$((unmatched + 1))
+        log "warning" "Skipped document with missing ID or name"
+        continue
+    fi
+
+    # Extract the first _YYYY.MM.DD_ pattern.
+    if [[ "$name" =~ (^|_)([0-9]{4})\.([0-9]{2})\.([0-9]{2})(_|$) ]]; then
+        year="${BASH_REMATCH[2]}"
+        month="${BASH_REMATCH[3]}"
+        day="${BASH_REMATCH[4]}"
+        document_date="$year-$month-$day"
+    else
+        unmatched=$((unmatched + 1))
+        printf '[SKIP] %s: date pattern not found\n' "$name"
+        continue
+    fi
+
+    # Reject impossible dates such as 2026.02.30.
+    parsed_date="$(
+        date \
+            --utc \
+            --date="$document_date" \
+            +"%Y-%m-%d" \
+            2>/dev/null ||
+            true
+    )"
+
+    if [[ "$parsed_date" != "$document_date" ]]; then
+        unmatched=$((unmatched + 1))
+        printf '[SKIP] %s: invalid date %s\n' \
+            "$name" \
+            "$document_date"
+        continue
+    fi
+
+    matched=$((matched + 1))
+
+    if [[ "$apply" != true ]]; then
+        printf '[DRY-RUN] %s -> %s\n' \
+            "$name" \
+            "$document_date"
+        continue
+    fi
+
+    payload="$(
+        jq \
+            --null-input \
+            --compact-output \
+            --arg documentDate "$document_date" \
+            '{documentDate: $documentDate}'
+    )"
+
+    if ! curl \
+        --fail \
+        --silent \
+        --show-error \
+        --request PATCH \
+        --header "Authorization: Bearer $TOKEN" \
+        --header "Content-Type: application/json" \
+        --data "$payload" \
+        --output /dev/null \
+        "$URL/api/organizations/$ORG_ID/documents/$id"
+    then
+        failed=$((failed + 1))
+        log "error" "Failed to update document: $id ($name)"
+        continue
+    fi
+
+    updated=$((updated + 1))
+
+    printf '[%d/%d] %s -> %s\n' \
+        "$updated" \
+        "$matched" \
+        "$name" \
+        "$document_date"
+done <"$documents_file"
+
+if [[ "$apply" == true ]]; then
+    log "info" "Update completed: read=$documents_read matched=$matched unmatched=$unmatched updated=$updated failed=$failed"
+else
+    log "info" "Dry run completed: read=$documents_read matched=$matched unmatched=$unmatched"
+    log "info" "No documents were modified. Run again with -a to apply"
+fi
+
+if (( failed > 0 )); then
+    exit 1
+fi

config/services/containers/app/papra/services/papra-backup.service

@@ -8,7 +8,7 @@ Type=oneshot
 # Create a consistent SQLite snapshot and atomically publish it.
 ExecStart=/usr/bin/podman exec -u nonroot papra node --input-type=module -e '\
         import { createClient } from "@libsql/client"; \
-        import { rename, rm } from "node:fs/promises"; \
+        import { rename, copyFile, rm } from "node:fs/promises"; \
         \
         /* Paths */ \
         const databaseUrl = "file:/app/app-data/db/db.sqlite"; \
@@ -22,7 +22,8 @@ ExecStart=/usr/bin/podman exec -u nonroot papra node --input-type=module -e '\
             String(now.getDate()).padStart(2, "0"), \
         ].join("-"); \
         \
-        const dst = backupDir + "/db_" + date + ".sqlite"; \
+        const dst = backupDir + "/db.sqlite"; \
+        const dated_dst = backupDir + "/db_" + date + ".sqlite"; \
         const tmp = dst + ".tmp"; \
         \
         /* Never expose a partially-created backup as a completed one. */ \
@@ -43,6 +44,7 @@ ExecStart=/usr/bin/podman exec -u nonroot papra node --input-type=module -e '\
             \
             /* Publish only after VACUUM INTO has completed successfully. */ \
             await rename(tmp, dst); \
+            await copyFile(dst, dated_dst); \
             completed = true; \
         } finally { \
             if (!completed) { \