Initial commit

.envrc

@@ -0,0 +1,11 @@
+#! /bin/sh
+
+# reload when these files change
+watch_file flake.nix
+watch_file flake.lock
+
+{
+  # shell gc root dir
+  mkdir -p "$(direnv_layout_dir)"
+  eval "$(nix print-dev-env --profile $(direnv_layout_dir)/flake-profile)"
+}

.gitignore

@@ -0,0 +1,4 @@
+.direnv/
+target/
+test/tmp/
+result

README.md

@@ -0,0 +1,59 @@
+# Docspell with audio files
+
+This is a simple addon for [Docspell](https://docspell.org) to add
+support for audio files. It uses external tools like ffmpeg and stt to
+extract text from an audio file. It uses wkhtmltopdf to create a pdf
+file.
+
+The result is a preview + pdf and indexed text in the Docspell dms.
+
+*Note: The accuracy of text extraction depends on the used models. For
+simplicty this addon uses pre-trained models. For better results, you
+need to train your own models.*
+
+
+## Prerequisites
+
+This addon supports these runners: `nix-flake` and `trivial`. An
+internet connection is required to download the model files.
+
+It is recommended to install [nix](https://nixos.org) on the machine
+running joex. This allows to use the `nix-flake` runner which can
+build the addon with all dependencies automatically.
+
+Otherwise, for the trivial runner, you need to install these tools
+manually: curl, ffmpeg, [stt](https://github.com/coqui-ai/STT)
+(v0.9.x) and wkhtmltopdf. The latter should already be available since
+it is a requirement for joex itself.
+
+Only `x86_64` archtiecture is supported, because nixpkgs doesn't
+provide `stt` for other architectures.
+
+
+## Usage
+
+Currently there is nothing to configure. Just install the addon and
+add it to a run configuration.
+
+
+## Testing
+
+Install [direnv](https://direnv.net/) and [nix](https://nixos.org) and
+allow the source root via `direnv allow`. This applies the `devShell`
+settings from `flake.nix`. Then build the addon:
+
+```
+nix build
+```
+
+Now you can run it:
+
+```
+./result/bin/audio-files-addon
+```
+
+It will run on the test files provided in `test/` and put results in
+`test/tmp`.
+
+For quicker turnaround you can also run the source file itself. This
+works, because `devShell` puts all required binaries in path.

docspell-addon.yml

@@ -0,0 +1,29 @@
+meta:
+  name: "audio-files-addon"
+  version: "0.1.0"
+  description: |
+    This addon adds support for audio files. Audio files are processed
+    by a speech-to-text engine and a pdf is generated.
+
+    It doesn't expect any user arguments at the moment. It requires
+    internet access to download model files.
+
+triggers:
+  - final-process-item
+  - final-reprocess-item
+  - existing-item
+
+runner:
+  nix:
+    enable: true
+
+  docker:
+    enable: false
+
+  trivial:
+    enable: true
+    exec: src/addon.scm
+
+options:
+  networking: true
+  collectOutput: true

flake.nix

@@ -0,0 +1,56 @@
+{
+  description = "A docspell addon for basic audio file support";
+
+  inputs = {
+    utils.url = "github:numtide/flake-utils";
+
+    # Nixpkgs / NixOS version to use.
+    nixpkgs.url = "nixpkgs/nixos-21.11";
+  };
+
+  outputs = { self, nixpkgs, utils }:
+    utils.lib.eachSystem ["x86_64-linux"] (system:
+      let
+        pkgs = import nixpkgs {
+          inherit system;
+          overlays = [
+
+          ];
+        };
+        name = "audio-files-addon";
+      in rec {
+        packages.${name} = pkgs.callPackage ./nix/addon.nix {
+          inherit name;
+        };
+
+        defaultPackage = packages.${name};
+
+        apps.${name} = utils.lib.mkApp {
+          inherit name;
+          drv = packages.${name};
+        };
+        defaultApp = apps.${name};
+
+        devShell = pkgs.mkShell {
+          inputsFrom = builtins.attrValues self.packages.${system};
+          buildInputs =
+            [ pkgs.guile
+              pkgs.guile-json
+              pkgs.stt
+              pkgs.wkhtmltopdf
+              pkgs.ffmpeg
+            ];
+
+          ADDON_DIR = self;
+          TMPDIR = "/tmp";
+          ITEM_DATA_JSON="test/item_data.json";
+          ITEM_ORIGINAL_JSON="test/item_source.json";
+          ITEM_ORIGINAL_DIR="test/original";
+          TMP_DIR="test/tmp";
+          CACHE_DIR="test/tmp";
+          OUTPUT_DIR="test/tmp";
+          GUILE_WARN_DEPRECATED="detailed";
+        };
+      }
+    );
+}

nix/addon.nix

@@ -0,0 +1,34 @@
+{ stdenv, bash, cacert, curl, stt, wkhtmltopdf, ffmpeg, guile, guile-json, lib, name }:
+
+stdenv.mkDerivation {
+  inherit name;
+  src = lib.sources.cleanSource ../.;
+
+  buildInputs = [ guile guile-json ];
+
+  patchPhase = ''
+    TARGET=src/addon.scm
+    sed -i 's,\*curl\* "curl",\*curl\* "${curl}/bin/curl",g' $TARGET
+    sed -i 's,\*ffmpeg\* "ffmpeg",\*ffmpeg\* "${ffmpeg}/bin/ffmpeg",g' $TARGET
+    sed -i 's,\*stt\* "stt",\*stt\* "${stt}/bin/stt",g' $TARGET
+    sed -i 's,\*wkhtmltopdf\* "wkhtmltopdf",\*wkhtmltopdf\* "${wkhtmltopdf}/bin/wkhtmltopdf",g' $TARGET
+  '';
+
+  buildPhase = ''
+    guild compile -o ${name}.go src/addon.scm
+  '';
+
+
+  # module name must be same as <filename>.go
+  installPhase = ''
+    mkdir -p $out/{bin,lib}
+    cp ${name}.go $out/lib/
+
+    cat > $out/bin/${name} <<-EOF
+    #!${bash}/bin/bash
+    export SSL_CERT_FILE="${cacert}/etc/ssl/certs/ca-bundle.crt"
+    exec -a "${name}" ${guile}/bin/guile -C ${guile-json}/share/guile/ccache -C $out/lib -e '(${name}) main' -c "" \$@
+    EOF
+    chmod +x $out/bin/${name}
+  '';
+}