Merge pull request #87 from sbs20/development

Airscan, multiple files -> zip etc
sbs20 · Oct 12, 2020 · e368ab1 · e368ab1
2 parents ce1f13c + 94b97ca
commit e368ab1
Show file tree

Hide file tree

Showing 14 changed files with 269 additions and 59 deletions.
diff --git a/Dockerfile b/Dockerfile
@@ -5,22 +5,38 @@ WORKDIR "$APP_DIR"
 COPY package*.json "$APP_DIR/"
 RUN npm install
 COPY . "$APP_DIR"
-RUN npm run server-build
-RUN npm run client-build
+RUN npm run server-build && npm run client-build
 
 # production image
 FROM node:buster-slim
 ENV APP_DIR=/app
 WORKDIR "$APP_DIR"
-# Install sane
-RUN apt-get update && apt-get install -yq sane sane-utils imagemagick tesseract-ocr
-RUN sed -i 's/policy domain="coder" rights="none" pattern="PDF"/policy domain="coder" rights="read | write" pattern="PDF"'/ /etc/ImageMagick-6/policy.xml
+RUN apt-get update && \
+  apt-get install -yq curl gpg && \
+  echo 'deb http://download.opensuse.org/repositories/home:/pzz/Debian_10/ /' | tee /etc/apt/sources.list.d/home:pzz.list && \
+  curl -fsSL https://download.opensuse.org/repositories/home:pzz/Debian_10/Release.key | gpg --dearmor | tee /etc/apt/trusted.gpg.d/home:pzz.gpg > /dev/null && \
+  apt-get update && \
+  apt-get install -yq sane sane-utils imagemagick tesseract-ocr sane-airscan && \
+  sed -i 's/policy domain="coder" rights="none" pattern="PDF"/policy domain="coder" rights="read | write" pattern="PDF"'/ /etc/ImageMagick-6/policy.xml
+
 COPY --from=builder "$APP_DIR/dist" "$APP_DIR/"
 
-# Install dependencies
 RUN npm install --production
 
-ENV NET_HOST=""
+# This goes into /etc/sane.d/net.conf
+ENV SANED_NET_HOSTS=""
+
+# This gets added to /etc/sane.d/airscan.conf
+ENV AIRSCAN_DEVICES=""
+
+# This directs scanserv not to bother querying `scanimage -L`
+ENV SCANIMAGE_LIST_IGNORE=""
+
+# This gets added to scanservjs/config/config.js:devices
+ENV DEVICES=""
+
+# Override OCR language
+ENV OCR_LANG=""
 
 # Copy entry point
 COPY run.sh /run.sh

diff --git a/README.md b/README.md
@@ -30,7 +30,7 @@ in these cases**
 ```console
 docker pull sbs20/scanservjs:latest
 docker rm --force scanservjs-container 2> /dev/null
-docker run -d -p 8080:8080 --restart unless-stopped --name scanservjs-container --privileged sbs20/scanservjs:latest
+docker run -d -p 8080:8080 -v /var/run/dbus:/var/run/dbus --restart unless-stopped --name scanservjs-container --privileged sbs20/scanservjs:latest
 ```
 (`--privileged` is required for the container to access the host's devices, to
 allow it to talk to the scanner)
@@ -42,7 +42,7 @@ If you want to install the latest staging branch (this may contain newer code)
 ```console
 docker pull sbs20/scanservjs:staging
 docker rm --force scanservjs-container 2> /dev/null
-docker run -d -p 8080:8080 --restart unless-stopped --name scanservjs-container --privileged sbs20/scanservjs:staging
+docker run -d -p 8080:8080 -v /var/run/dbus:/var/run/dbus --restart unless-stopped --name scanservjs-container --privileged sbs20/scanservjs:staging
 ```
 
 More installation options:
@@ -51,6 +51,51 @@ More installation options:
 * [Development notes](docs/development.md)
 * [Configuring the scanner and SANE](docs/sane.md)
 
+## Environment variables
+
+* `SANED_NET_HOSTS`: If you want to use a
+  [SaneOverNetwork](https://wiki.debian.org/SaneOverNetwork#Server_Configuration)
+  scanner then to perform the equivalent of adding hosts to
+  `/etc/sane.d/net.conf` specify a list of ip addresses separated by semicolons
+  in the `SANED_NET_HOSTS` environment variable.
+* `AIRSCAN_DEVICES`: If you want to specifically add `sane-airscan` devices to
+  your `/etc/sane.d/airscan.conf` then use the `AIRSCAN_DEVICES` environment
+  variable (semicolon delimited).
+* `DEVICES`: Force add devices use `DEVICES` (semicolon delimited)
+* `SCANIMAGE_LIST_IGNORE`: To force ignore `scanimage -L`
+
+## Airscan
+[sane-airscan](https://github.com/alexpevzner/sane-airscan) uses Avahi /
+Zeroconf / Bonjour to discover devices on the local network. If you are running
+docker you will want to share dbus to make it work
+(`-v /var/run/dbus:/var/run/dbus`).
+
+## Example docker run
+
+### Use airscan and a locally detected scanner
+This should support most use cases
+
+```console
+docker run -d -p 8080:8080 \
+  -v /var/run/dbus:/var/run/dbus \
+  --name scanservjs-container --privileged scanservjs-image
+```
+
+### Complicated
+Add two net hosts to sane, use airscan to connect to two remote scanners, don't
+use `scanimage -L`, force a list of devices and override the OCR language
+
+```console
+docker run -d -p 8080:8080 \
+  -e SANED_HOSTS="10.0.100.30;10.0.100.31" \
+  -e AIRSCAN_DEVICES='"Canon MFD" = "http://192.168.0.10/eSCL";"EPSON MFD" = "http://192.168.0.11/eSCL"' \
+  -e SCANIMAGE_LIST_IGNORE=true \
+  -e DEVICES="net:10.0.100.30:plustek:libusb:001:003;net:10.0.100.31:plustek:libusb:001:003;airscan:e0:Canon TR8500 series;airscan:e1:EPSON Cool Series" \
+  -e OCR_LANG="fra" \
+  -v /var/run/dbus:/var/run/dbus \
+  --name scanservjs-container --privileged scanservjs-image
+```
+
 ## Why?
 This is yet another scanimage-web-front-end. Why? It originally started as an
 adaptation of phpsane - just to make everything a bit newer, give it a refresh

diff --git a/client/components/Scanserv.vue b/client/components/Scanserv.vue
@@ -381,13 +381,21 @@ export default {
 
       return this._fetch(url).then(context => {
         this.context = context;
-        this.device = context.devices[0];
-        this.$refs.toastr.i(`Found device ${this.device.id}`);
-        this.request = this.readRequest();
-        for (let test of context.diagnostics) {
-          const toast = test.success ? this.$refs.toastr.s : this.$refs.toastr.e;
-          toast(test.message);
+
+        if (context.devices.length > 0) {
+          for (let device of context.devices) {
+            this.$refs.toastr.i(`Found device ${device.id}`);
+          }
+          this.device = context.devices[0];
+          this.request = this.readRequest();
+          for (let test of context.diagnostics) {
+            const toast = test.success ? this.$refs.toastr.s : this.$refs.toastr.e;
+            toast(test.message);
+          }
+        } else {
+          this.$refs.toastr.e('Found no devices');
         }
+
         if (force) {
           this.clear();
           this.readPreview();
@@ -483,7 +491,7 @@ export default {
         }
       }).then((data) => {
         if (data && 'page' in data) {
-          if (window.confirm('Scan another page?')) {
+          if (window.confirm(`Scan page ${data.page}?`)) {
             this.request.page = data.page;
             this.scan();
           } else {

diff --git a/config/config.js b/config/config.js
@@ -4,10 +4,6 @@ const Config = {};
 
 // Things to change
 Config.port = 8080;
-
-// scanservjs will attempt to find scanners locally using `scanimage -L` but you
-// will need to manually add network devices here which will be appended. e.g.
-// Config.devices = ['net:192.168.0.10:airscan:e0:Canon TR8500 series'];
 Config.devices = [];
 Config.ocrLanguage = 'eng';
 Config.log = {};
@@ -46,29 +42,38 @@ Config.previewPipeline = {
   ]
 };
 
-/*
-When all scans are complete, the filenames are all piped into stdin to the
-pipeline commands. It would be nicer to pipe the binary output of scanimage but
-that doesn't work with multipage scans so we have no choice but to write to the
-filesystem.
+/* When all scans are complete, the filenames are all piped into stdin of the
+first pipeline command. It would be nicer to pipe the binary output of scanimage
+but that doesn't work with multipage scans so we have no choice but to write to
+the filesystem.
 
 The stdout of each pipeline feeds into the stdin of the next. Although clumsy in
 some respects (especially where we have to write temporary files and then list
 them) it at least provides a means of user configuration with "just" shell
 scripting.
 
+The overall output of the pipelines (i.e. the last pipeline output) must be a
+list of the files you want kept. The convention is to output files of the form
+`scan-0000.ext` but it's convention only. You can output whatever you want. If
+multiple files are output then the results will be zipped into a single file.
+
 Each command is executed with the CWD set to the temporary location so no
 directory traversal is required. Pipeline commands are always read from this
 file (and never from the browser request, even though it is sent). It would be
 possible to subvert these commands for malicious use, but it doesn't give any
-further privilege than the user account running scanservjs. You obviously should
-not be running as root.
+further privilege than the user account running scanservjs and still requires
+access to this file. You obviously should not be running as root.
 
 Some useful pointers:
 - `convert` can read a list of files from a file with the @ argument. The `-`
-  file is stdin. So `convert @- -argument output` performs the conversion om
+  file is stdin. So `convert @- -argument output` performs the conversion on
   each file piped into stdin
 - `tesseract` has a similar feature using `-c stream_filelist=true`
+- `convert` can also output multiple files if you use an output filename with
+  `%d` in it. C string style formatting is available so you can do things like
+  output to `scan-%04d.jpg`. Formats which do not support multiple pages must
+  use this option. Multi-page formats including PDF and TIF do not use this
+  option.
 - if you just wanted to take a filename from stdin and have its content read out
   you could `xargs cat` provided there were no spaces or commas in the filename
   (which there won't be)
@@ -78,81 +83,92 @@ Config.pipelines = [
     extension: 'jpg',
     description: 'JPG | High quality',
     commands: [
-      'convert @- -quality 92 jpg:-'
+      'convert @- -quality 92 scan-%04d.jpg',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'jpg',
     description: 'JPG | Medium quality',
     commands: [
-      'convert @- -quality 75 jpg:-'
+      'convert @- -quality 75 scan-%04d.jpg',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'jpg',
     description: 'JPG | Low quality',
     commands: [
-      'convert @- -quality 50 jpg:-'
+      'convert @- -quality 50 scan-%04d.jpg',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'png',
     description: 'PNG',
     commands: [
-      'convert @- -quality 75 png:-'
+      'convert @- -quality 75 scan-%04d.png',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'tif',
     description: 'TIF | Uncompressed',
     commands: [
-      'convert @- tif:-'
+      'convert @- scan-0000.tif',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'tif',
     description: 'TIF | LZW compression',
     commands: [
-      'convert @- -compress lzw tif:-'
+      'convert @- -compress lzw scan-0000.tif',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'pdf',
     description: 'PDF (TIF | Uncompressed)',
     commands: [
-      'convert @- pdf:-'
+      'convert @- scan-0000.pdf',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'pdf',
     description: 'PDF (TIF | LZW Compression)',
     commands: [
-      'convert @- -compress lzw tmp-%d.tif && ls tmp-*.tif',
-      'convert @- pdf:-'
+      'convert @- -compress lzw tmp-%04d.tif && ls tmp-*.tif',
+      'convert @- scan-0000.pdf',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'pdf',
     description: 'PDF (JPG | High quality)',
     commands: [
-      'convert @- -quality 92 tmp-%d.jpg && ls tmp-*.jpg',
-      'convert @- pdf:-'
+      'convert @- -quality 92 tmp-%04d.jpg && ls tmp-*.jpg',
+      'convert @- scan-0000.pdf',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'pdf',
     description: 'PDF (JPG | Medium quality)',
     commands: [
-      'convert @- -quality 75 tmp-%d.jpg && ls tmp-*.jpg',
-      'convert @- pdf:-'
+      'convert @- -quality 75 tmp-%04d.jpg && ls tmp-*.jpg',
+      'convert @- scan-0000.pdf',
+      'ls scan-*.*'
     ]
   },
   {
     extension: 'pdf',
     description: 'PDF (JPG | Low quality)',
     commands: [
-      'convert @- -quality 50 tmp-%d.jpg && ls tmp-*.jpg',
-      'convert @- pdf:-'
+      'convert @- -quality 50 tmp-%04d.jpg && ls tmp-*.jpg',
+      'convert @- scan-0000.pdf',
+      'ls scan-*.*'
     ]
   }
 ];
@@ -164,17 +180,42 @@ if (Config.tesseract) {
       description: 'OCR | PDF (JPG | High quality)',
       commands: [
         'convert @- -quality 92 tmp-%d.jpg && ls tmp-*.jpg',
-        `${Config.tesseract} -l ${Config.ocrLanguage} -c stream_filelist=true - - pdf && rm -f tmp-*.jpg`
+        `${Config.tesseract} -l ${Config.ocrLanguage} -c stream_filelist=true - - pdf > scan-0001.pdf`,
+        'ls scan-*.*'
       ]
     },
     {
       extension: 'txt',
       description: 'OCR | Text file',
       commands: [
-        `${Config.tesseract} -l ${Config.ocrLanguage} -c stream_filelist=true - - txt && rm -f tmp-*.tif`
+        `${Config.tesseract} -l ${Config.ocrLanguage} -c stream_filelist=true - - txt > scan-0001.txt`,
+        'ls scan-*.*'
       ]
     }
   ]);
 }
 
+// Process environment variables
+
+// scanservjs will attempt to find scanners locally using `scanimage -L` but
+// sometimes you may need to manually add network devices here if they're not
+// found e.g.
+// Config.devices = ['net:192.168.0.10:airscan:e0:Canon TR8500 series'];
+// This is done with an environment variable. Multiple entries are separated by
+// semicolons
+if (process.env.DEVICES !== undefined && process.env.DEVICES.length > 0) {
+  Config.devices = process.env.DEVICES.split(';');
+}
+
+// scanservjs will attempt to find scanners locally using `scanimage -L` but
+// sometimes it will return nothing. If you are specifying devices manually you
+// may also with to turn off the find.
+Config.devicesFind = process.env.SCANIMAGE_LIST_IGNORE === undefined
+  || process.env.SCANIMAGE_LIST_IGNORE.length === 0;
+
+// Override the OCR language here
+if (process.env.OCR_LANG !== undefined && process.env.OCR_LANG.length > 0) {
+  Config.ocrLanguage = process.env.OCR_LANG;
+}
+
 module.exports = Config;
diff --git a/docs/development.md b/docs/development.md
@@ -54,6 +54,10 @@ gulp release
 Install docker
 ```
 sudo apt install docker.io
+sudo systemctl unmask docker
+sudo systemctl start docker
+
+# Hack to make docker accessible.
 sudo chmod 666 /var/run/docker.sock
 ```
 
@@ -62,7 +66,11 @@ Useful commands
 # Build and run
 docker build -t scanservjs-image .
 docker rm --force scanservjs-container 2> /dev/null
-docker run -d -p 8080:8080 --name scanservjs-container --privileged scanservjs-image
+docker run -d -p 8080:8080 -v /var/run/dbus:/var/run/dbus --name scanservjs-container --privileged scanservjs-image
+
+# Copy image
+docker save -o scanservjs-image.tar scanservjs-image
+docker load -i scanservjs-image.tar
 
 # Debug
 docker run -it --entrypoint=/bin/bash scanservjs-container