setup auto scan folder and ocr new pdf

[BeckeBauer]

First of all, great app 👍
I just noticed the option Local Directory Scanning and File Processing and tried to set it up. Since I am running the app via docker, I mounted my local scan folder under ./pipeline/watchedFolders.

However, I am lost when it comes to creating a JSON configuration file. I used the following configuration which I was able to extract from the pipeline feature:

{
  "name": "ocr",
  "pipeline": [
    {
      "operation": "/api/v1/misc/ocr-pdf",
      "parameters": {
        "fileInput": "automated"
      }
    }
  ],
  "_examples": {
    "outputDir": "{outputFolder}/{folderName}",
    "outputFileName": "{filename}-{pipelineName}-{date}-{time}"
  },
  "outputDir": "{outputFolder}",
  "outputFileName": "{filename}"
}

Unfortunately, clicking on the wheel with yellow background to configure the pipeline settings does not do anything in my case, so I have no idea how the parameter (in particular changing ocr language) could look like.

Also, how can I define that the "ocr-ed" files are saved in the same directory they are in?

Is there anything that needs to be added to the configuration?

Any sample JSON configuration file or help is appreciated

[Frooodle]

seems to be a bug in latest release, Will try patch soon!

[BeckeBauer]

Okay, great but I am kind of confused. Do you refer to the issue that the settings wheel in the pipeline feature is not working?

If so, I was wondering if you could help me with the creation of the JSON config file (configuration of the output folder)?

[Frooodle]

please update and try recreate the config!

[BeckeBauer]

After updating, JSON config files can be created with the pipeline feature 👍. For ocr, I got the following config:

{
  "name": "ocr",
  "pipeline": [
    {
      "operation": "/api/v1/misc/ocr-pdf",
      "parameters": {
        "languages": "deu, eng",
        "sidecar": false,
        "deskew": false,
        "clean": false,
        "cleanFinal": false,
        "ocrType": "skip-text",
        "ocrRenderType": "hocr",
        "removeImagesAfter": false,
        "fileInput": "automated"
      }
    }
  ],
  "_examples": {
    "outputDir": "{outputFolder}/{folderName}",
    "outputFileName": "{filename}-{pipelineName}-{date}-{time}"
  },
  "outputDir": "{outputFolder}",
  "outputFileName": "{filename}"
}

The configuration works when I use it to ocr pdf-files in the browser. So, I saved the ocr.json in the folder ./pipeline/watchedFolders/Scans and put some pdf-files into the folder but nothing happens.

Is there anything I need to do to activate the feature Local Directory Scanning and File Processing?

[Frooodle]

Weird that should work
Let me try reproduce for you

[Frooodle]

One more fix coming, will be released in about 10 minutes, 0.22.6

[Frooodle]

If you update it should start automatically working now :)

[BeckeBauer]

Now it works, thank you 👌

[BeckeBauer]

If someone is interested: I created two subfolders under my folder Scan (scan_in and scan_ocr) and mounted both folders via docker

- /path to/Scans/scans_in:/pipeline/watchedFolders/scans_in 
- /path to/Scans/scans_ocr:/pipeline/finishedFolders/scans_ocr

If you now place your pdfs into the scan_in folder and wait a bit, you can find them in scans_ocr as searchable pdf after some time