added mgw

README.md

@@ -32,18 +32,14 @@ The `generate` command accepts one input file or one input folder and one output
 #### Usage
 
 ```
-generate [-F|--format] [-c|--config <config-file>] <input-file> <output-directory>
+generate <input-file> <output-directory>
     <input-file>                The file to generate embeddings for
     <output-directory>          The directory to write the embeddings to
     -F|--format [csv|parquet]   The format of the embeddings file: `CSV`` or `Parquet``
     -c|--config <config-file>   An optional configuration file to use
 ```
 
-#### Configuration file
 
-```yaml
-hop_length_seconds: 5.0
-```
 
 #### Examples
 
@@ -75,60 +71,9 @@ classify [-F|--format] [-c|--config <config-file> ] <input-file> <output-directo
 #### Configuration file
 
 ```yaml
-hop_length_seconds: 5.0
-classifiers:
-  - name: "classifier1"
-    tags: ["tag1", "tag2"]
-    # base64 encoded model
-    model: "asd321jhfsdhk438asdmlfas89i3mnlksf..."
-  - name: "classifier2"
-    tags: ["tag3", "tag4"]
-     # base64 encoded model
-    model: "..."
-
-```
-
-#### Example
-
-```bash 
-cd /data
-curl -o audio.wav https://api.ecosounds.org/audio_recordings/123.wav
-docker run -v /data:/data perch-runner classify -c /data/config.yml /data/audio.wav /data/output
-```
-
-### Distance
-
-The `distance` command accepts one input file, one output directory, and a query file.
-
-- The query file must be an audio file that is 5 seconds long.
-- The output of this command is the same as the output of the `generate` command,
-    but with an additional column `distance` that represents the distance between the
-    query and the sample.
-
-#### Usage
-
-```
-distance [-F|--format] [-c|--config <config-file> ] <input-file> <query-file> <output-directory> 
-    <input-file>                The file to generate embeddings for
-    <output-directory>          The directory to write the embeddings to
-    <query-file>                The file to use as the query
-    -c|--config <config-file>   An optional configuration file to use
-```
-
-#### Configuration file
-
-```yaml
-hop_length_seconds: 5.0
+classifier: keras_saved_model_path
 ```
 
-#### Example
-
-```bash
-cd /data
-curl -o audio.wav https://api.ecosounds.org/audio_recordings/123.wav
-curl -o query.wav https://api.ecosounds.org/audio_recordings/456/media.wav?start=10&end=15
-docker run -v /data:/data perch-runner distance -c /data/config.yml /data/audio.wav /data/query.wav /data/output
-```
 
 
 ## Batch Mode

models/mgw/mgw_v05.keras.labels.json

@@ -0,0 +1 @@
+["mgw", "negative"]

scripts/Readme.md

@@ -25,8 +25,11 @@ Notes
 1. Open a terminal window
 2. Change directory to this scripts directory
 3. Run the following command:
-  - windows: `powershell -ExecutionPolicy Bypass -File analyze.ps1 classify [path_to_audio_folder] [path_to_embeddings_folder] 'pw'`
-  - linux or intel mac: `./analyze.sh classify [path_to_embeddings_output_folder] [path_to_classifications_output_folder] 'pw'`
+  - windows: `powershell -ExecutionPolicy Bypass -File analyze.ps1 classify [path_to_audio_folder] [path_to_embeddings_folder] '[recognizer_id]'`
+  - linux or intel mac: `./analyze.sh classify [path_to_embeddings_output_folder] [path_to_classifications_output_folder] '[recognizer_id]'`
+  - where [recognizer_id] is one of
+    - 'pw' (Plains wanderer)
+    - 'mgw' (Mukarrthippi graw wren)
 
 
 Notes

scripts/analyze.ps1

@@ -39,10 +39,11 @@ $config = ""
 $recognizer_configs = @{
     "pw" = "pw.classify.yml"
     "cgw" = "cgw.classify.yml"
+    "mgw" = "mgw.classify.yml"
 }
 
 # Check if the recognizer argument has been provided
-if ($recognizer) {
+if ($recognizer -and $analysis -eq "classify") {
     # Check if the provided recognizer is supported and set the config variable
     if ($recognizer_configs.ContainsKey($recognizer)) {
         $config = $recognizer_configs[$recognizer]

scripts/analyze.sh

@@ -35,27 +35,33 @@ fi
 # Initialize command with a default value
 command=""
 
-declare -A recognizer_configs
-recognizer_configs["pw"]="pw.classify.yml"
-recognizer_configs["cgw"]="cgw.classify.yml"
-
-# Check if the recognizer argument has been provided
-if [[ -n "$recognizer" ]]; then
-    # Check if the provided recognizer is supported and set the config variable
-    if [[ -n ${recognizer_configs[$recognizer]} ]]; then
-        config=${recognizer_configs[$recognizer]}
-        echo "Using config file: $config"
-        config="--config $config"
+# if analysis is 'classify'
+if [[ "$analysis" == "classify" ]]; then
+    declare -A recognizer_configs
+    recognizer_configs["pw"]="pw.classify.yml"
+    recognizer_configs["cgw"]="cgw.classify.yml"
+    recognizer_configs["mgw"]="mgw.classify.yml"
+
+    # Check if the recognizer argument has been provided
+    if [[ -n "$recognizer" ]]; then
+        # Check if the provided recognizer is supported and set the config variable
+        if [[ -n ${recognizer_configs[$recognizer]} ]]; then
+            config=${recognizer_configs[$recognizer]}
+            echo "Using config file: $config"
+            config="--config $config"
+        else
+            echo "Recognizer $recognizer not supported"
+            exit 1
+        fi
     else
-        echo "Recognizer $recognizer not supported"
-        exit 1
+        # Set config variable to an empty string if no recognizer is provided
+        config=""
     fi
-else
-    # Set config variable to an empty string if no recognizer is provided
-    config=""
 fi
 
 
+
+
 # paths to things inside the container, to be mounted
 
 # Determine input container path

src/default_configs/mgw.classify.yml

@@ -0,0 +1,2 @@
+classifier: mgw
+inherit: classify.yml

tests/docker_tests/test_analyze.py

@@ -7,7 +7,8 @@
     ["pwsh", "-ExecutionPolicy", "Bypass", "-File",  "./scripts/analyze.ps1"]
 ]
 
-index = 1
+index = 0
+#image = "pr10:latest"
 
 def test_analyze_script_classify():
 
@@ -19,7 +20,7 @@ def test_analyze_script_classify():
     shutil.copy("tests/files/embeddings/100sec.parquet", two)
 
     # Run the embed helper script
-    command = scripts[index] + ["classify", "./tests/input/", "./tests/output/", "pw"]
+    command = scripts[index] + ["classify", "./tests/input/", "./tests/output/", "mgw"]
     print(f'running command: {" ".join(command)}')
     subprocess.run(command, check=True)