Add Italian Kaldi model

.gitmodules

@@ -105,3 +105,6 @@
 [submodule "czech/cs_kaldi-rhasspy"]
 	path = czech/cs_kaldi-rhasspy
 	url = https://github.com/rhasspy/cs_kaldi-rhasspy.git
+[submodule "italian/it_kaldi-rhasspy"]
+	path = italian/it_kaldi-rhasspy
+	url = https://github.com/rhasspy/it_kaldi-rhasspy.git

PROFILES

@@ -1,4 +1,5 @@
 catalan/ca-es_pocketsphinx-cmu
+czech/cs_kaldi-rhasspy
 dutch/nl_kaldi-cgn
 dutch/nl_pocketsphinx-cmu
 english/en-in_pocketsphinx-cmu
@@ -13,6 +14,7 @@ german/de_kaldi-zamia
 german/de_pocketsphinx-cmu
 greek/el-gr_pocketsphinx-cmu
 hindi/hi_pocketsphinx-cmu
+italian/it_kaldi-rhasspy
 italian/it_pocketsphinx-cmu
 kazakh/kz_pocketsphinx-cmu
 korean/ko-kr_kaldi-montreal

README.md

@@ -225,6 +225,16 @@ Untested profiles (highlighted below) *may* work, but I don't have the necessary
       <td>&#9733; &#9733; &#9733; &#9733; &#9733; (21x)</td>
       <td>&#9733; &#9733; &#9733; &#9733; &#9733; (7x)</td>
     </tr>
+    <tr>
+      <td>
+        <a href="https://github.com/synesthesiam/it_kaldi-rhasspy/archive/v1.0.tar.gz">Download</a>
+      </td>
+      <td>Italian (Italiano)</td>
+      <td>it</td>
+      <td>kaldi</td>
+      <td>&#9733; &#9733; &#9733; &#9733; &#9733; (1x)</td>
+      <td>&#9733; &#9733; &#9733; &#9733; &#9733; (1x)</td>
+    </tr>
     <tr>
       <td>
         <a href="https://github.com/synesthesiam/kz_pocketsphinx-cmu/archive/v1.1.tar.gz">Download</a>

__main__.py

@@ -29,6 +29,10 @@
 @app.route("/<path:path>")
 async def download_raw(path: str) -> Response:
     components = path.split("/")
+
+    if components[0] in ("synesthesiam", "rhasspy"):
+        components = components[1:]
+
     profile = components[0]
 
     if components[1] == "raw":

bin/make_g2p.sh

@@ -28,6 +28,14 @@ trap finish EXIT
 
 # -----------------------------------------------------------------------------
 
+if [[ "$1" == '-' ]]; then
+    # Read from stdin into temporary file
+    dict_path="${temp_dir}/unformatted.dict"
+    cat > "${dict_path}"
+else
+    dict_path="$(realpath "$1")"
+fi
+
 # Format dictionary for phonetisaurus
 cd "${temp_dir}"
 perl -pe 's/\([0-9]+\)//;

bin/print_phonemes.py

@@ -2,6 +2,7 @@
 import sys
 import re
 import argparse
+from collections import Counter
 
 
 def main():
@@ -10,21 +11,41 @@ def main():
         description="Prints the unique set of phonemes from a pronunciation dictionary",
     )
     parser.add_argument("dictionary", help="Path to pronunciation dictionary")
+    parser.add_argument(
+        "--min-count",
+        type=int,
+        help="Minimum use count before printing phoneme in list",
+    )
+    parser.add_argument(
+        "--counts", action="store_true", help="Print counts with list"
+    )
     args = parser.parse_args()
 
-    phonemes = set()
-    with open(args.dictionary, "r") as dict_file:
+    if args.dictionary == "-":
+        dict_file = sys.stdin
+    else:
+        dict_file = open(args.dictionary, "r")
+
+    phonemes = Counter()
+    with dict_file:
         for line in dict_file:
             line = line.strip()
             if len(line) == 0:
                 continue
 
             # Use explicit whitespace (avoid 0xA0)
             parts = re.split(r"[ \t]+", line)
-            phonemes.update(parts[1:])
 
-        for phoneme in sorted(phonemes):
-            print(phoneme)
+            for phoneme in parts[1:]:
+                phonemes[phoneme] += 1
+
+    for phoneme in sorted(phonemes):
+        count = phonemes[phoneme]
+        if args.min_count is None or (count >= args.min_count):
+            if args.counts:
+                print(phoneme, count)
+            else:
+                print(phoneme)
 
 
 # -----------------------------------------------------------------------------