From 6e105fe52b61458c159ffd616bed36ef959d207e Mon Sep 17 00:00:00 2001
From: karamouche <hugo4ibt@gmail.com>
Date: Wed, 15 Apr 2026 14:45:29 -0400
Subject: [PATCH 1/4] feat: add CLI for text normalization with language and
 preset options

---
 README.md                 | 22 ++++++++++++
 normalization/__main__.py |  3 ++
 normalization/cli.py      | 73 +++++++++++++++++++++++++++++++++++++++
 pyproject.toml            |  3 ++
 4 files changed, 101 insertions(+)
 create mode 100644 normalization/__main__.py
 create mode 100644 normalization/cli.py

diff --git a/README.md b/README.md
index 6231cdb..96fc928 100644
--- a/README.md
+++ b/README.md
@@ -71,6 +71,28 @@ pipeline.normalize("It's $50 at 3:00PM")
 # => "it is 50 dollars at 3 pm"
 ```
 
+### CLI
+
+```bash
+# Normalize a single text
+normalize "It's $50 at 3:00PM" --language en
+
+# Pipe from stdin
+echo "she spent twenty dollars" | normalize --language fr
+
+# Use a custom preset
+normalize "some text" --preset path/to/my-preset.yaml --language en
+
+# Inspect the pipeline
+normalize --describe --language en
+```
+
+If you don't want a permanent installation, run it directly with `uvx`:
+
+```bash
+uvx gladia-normalization "It's $50 at 3:00PM" --language en
+```
+
 ## How it works
 
 Every pipeline runs exactly **three stages**, always in this order:
diff --git a/normalization/__main__.py b/normalization/__main__.py
new file mode 100644
index 0000000..5c728b0
--- /dev/null
+++ b/normalization/__main__.py
@@ -0,0 +1,3 @@
+from normalization.cli import main
+
+main()
diff --git a/normalization/cli.py b/normalization/cli.py
new file mode 100644
index 0000000..67a44ed
--- /dev/null
+++ b/normalization/cli.py
@@ -0,0 +1,73 @@
+import argparse
+import json
+import sys
+
+from normalization import load_pipeline
+from normalization.languages import get_language_registry
+from normalization.pipeline.loader import _PRESETS_DIR
+
+
+def _available_languages() -> list[str]:
+    import normalization.languages  # noqa: F401 — triggers @register_language decorators
+
+    return sorted(get_language_registry().keys())
+
+
+def _available_presets() -> list[str]:
+    return sorted(p.stem for p in _PRESETS_DIR.glob("*.yaml"))
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(
+        prog="normalize",
+        description="Normalize STT transcription text for fair WER comparison.",
+    )
+    parser.add_argument(
+        "text",
+        nargs="?",
+        help="Text to normalize. Reads from stdin if omitted.",
+    )
+    parser.add_argument(
+        "--language",
+        "-l",
+        default="en",
+        metavar="CODE",
+        help="Language code (default: en). Available: %(choices)s.",
+        choices=_available_languages(),
+    )
+    parser.add_argument(
+        "--preset",
+        "-p",
+        default="gladia-3",
+        metavar="PRESET",
+        help="Built-in preset name or path to a YAML file (default: gladia-3).",
+    )
+    parser.add_argument(
+        "--describe",
+        action="store_true",
+        help="Print the pipeline description as JSON and exit.",
+    )
+
+    args = parser.parse_args()
+
+    try:
+        pipeline = load_pipeline(args.preset, args.language)
+    except FileNotFoundError as exc:
+        parser.error(str(exc))
+
+    if args.describe:
+        print(json.dumps(pipeline.describe(), indent=2))
+        return
+
+    if args.text is not None:
+        text = args.text
+    elif not sys.stdin.isatty():
+        text = sys.stdin.read().strip()
+    else:
+        parser.error("Provide text as an argument or pipe it via stdin.")
+
+    print(pipeline.normalize(text))
+
+
+if __name__ == "__main__":
+    main()
diff --git a/pyproject.toml b/pyproject.toml
index fcae854..267c571 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -33,6 +33,9 @@ dependencies = [
 [tool.setuptools.package-data]
 normalization = ["presets/*.yaml"]
 
+[project.scripts]
+normalize = "normalization.cli:main"
+
 [project.urls]
 Homepage = "https://github.com/gladiaio/normalization"
 Repository = "https://github.com/gladiaio/normalization"

From e032c2a58713abcba153e57c06f5fcbae8da58f4 Mon Sep 17 00:00:00 2001
From: karamouche <hugo4ibt@gmail.com>
Date: Wed, 15 Apr 2026 15:37:55 -0400
Subject: [PATCH 2/4] feat: add support for normalizing text files via CLI

---
 README.md            |  3 +++
 normalization/cli.py | 19 +++++++++++++++++--
 2 files changed, 20 insertions(+), 2 deletions(-)

diff --git a/README.md b/README.md
index 96fc928..4d4afeb 100644
--- a/README.md
+++ b/README.md
@@ -77,6 +77,9 @@ pipeline.normalize("It's $50 at 3:00PM")
 # Normalize a single text
 normalize "It's $50 at 3:00PM" --language en
 
+# Normalize a text file
+normalize --file transcript.txt --language en
+
 # Pipe from stdin
 echo "she spent twenty dollars" | normalize --language fr
 
diff --git a/normalization/cli.py b/normalization/cli.py
index 67a44ed..cf04399 100644
--- a/normalization/cli.py
+++ b/normalization/cli.py
@@ -42,6 +42,12 @@ def main() -> None:
         metavar="PRESET",
         help="Built-in preset name or path to a YAML file (default: gladia-3).",
     )
+    parser.add_argument(
+        "--file",
+        "-f",
+        metavar="PATH",
+        help="Path to a text file to normalize. Mutually exclusive with positional text.",
+    )
     parser.add_argument(
         "--describe",
         action="store_true",
@@ -50,6 +56,9 @@ def main() -> None:
 
     args = parser.parse_args()
 
+    if args.text is not None and args.file is not None:
+        parser.error("Provide either positional text or --file, not both.")
+
     try:
         pipeline = load_pipeline(args.preset, args.language)
     except FileNotFoundError as exc:
@@ -59,12 +68,18 @@ def main() -> None:
         print(json.dumps(pipeline.describe(), indent=2))
         return
 
-    if args.text is not None:
+    if args.file is not None:
+        try:
+            with open(args.file) as fh:
+                text = fh.read().strip()
+        except OSError as exc:
+            parser.error(str(exc))
+    elif args.text is not None:
         text = args.text
     elif not sys.stdin.isatty():
         text = sys.stdin.read().strip()
     else:
-        parser.error("Provide text as an argument or pipe it via stdin.")
+        parser.error("Provide text as an argument, --file, or pipe it via stdin.")
 
     print(pipeline.normalize(text))
 

From ca5947de65a3c2e08e210718cafb539d70d61fea Mon Sep 17 00:00:00 2001
From: karamouche <hugo4ibt@gmail.com>
Date: Thu, 16 Apr 2026 08:53:49 -0400
Subject: [PATCH 3/4] refactor: rename CLI command to 'gladia-normalization'
 for consistency

---
 README.md      | 10 +++++-----
 pyproject.toml |  2 +-
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/README.md b/README.md
index 4d4afeb..53798c9 100644
--- a/README.md
+++ b/README.md
@@ -75,19 +75,19 @@ pipeline.normalize("It's $50 at 3:00PM")
 
 ```bash
 # Normalize a single text
-normalize "It's $50 at 3:00PM" --language en
+gladia-normalization "It's $50 at 3:00PM" --language en
 
 # Normalize a text file
-normalize --file transcript.txt --language en
+gladia-normalization --file transcript.txt --language en
 
 # Pipe from stdin
-echo "she spent twenty dollars" | normalize --language fr
+echo "she spent twenty dollars" | gladia-normalization --language fr
 
 # Use a custom preset
-normalize "some text" --preset path/to/my-preset.yaml --language en
+gladia-normalization "some text" --preset path/to/my-preset.yaml --language en
 
 # Inspect the pipeline
-normalize --describe --language en
+gladia-normalization --describe --language en
 ```
 
 If you don't want a permanent installation, run it directly with `uvx`:
diff --git a/pyproject.toml b/pyproject.toml
index 267c571..eac2b9c 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -34,7 +34,7 @@ dependencies = [
 normalization = ["presets/*.yaml"]
 
 [project.scripts]
-normalize = "normalization.cli:main"
+gladia-normalization = "normalization.cli:main"
 
 [project.urls]
 Homepage = "https://github.com/gladiaio/normalization"

From f5288f87bc9597840e3c61efc4dab8ee2ab7f317 Mon Sep 17 00:00:00 2001
From: karamouche <hugo4ibt@gmail.com>
Date: Thu, 16 Apr 2026 09:03:56 -0400
Subject: [PATCH 4/4] docs: update examples in README to avoid using $

---
 README.md | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/README.md b/README.md
index 53798c9..bca2733 100644
--- a/README.md
+++ b/README.md
@@ -75,16 +75,16 @@ pipeline.normalize("It's $50 at 3:00PM")
 
 ```bash
 # Normalize a single text
-gladia-normalization "It's $50 at 3:00PM" --language en
+gladia-normalization "A tea cost £2 at 3:30PM" --language en
 
 # Normalize a text file
 gladia-normalization --file transcript.txt --language en
 
 # Pipe from stdin
-echo "she spent twenty dollars" | gladia-normalization --language fr
+echo "A tea cost £2 at 3:30PM" | gladia-normalization --language fr
 
 # Use a custom preset
-gladia-normalization "some text" --preset path/to/my-preset.yaml --language en
+gladia-normalization "A tea cost £2 at 3:30PM" --preset path/to/my-preset.yaml --language en
 
 # Inspect the pipeline
 gladia-normalization --describe --language en
@@ -93,7 +93,7 @@ gladia-normalization --describe --language en
 If you don't want a permanent installation, run it directly with `uvx`:
 
 ```bash
-uvx gladia-normalization "It's $50 at 3:00PM" --language en
+uvx gladia-normalization "A tea cost £2 at 3:30PM" --language en
 ```
 
 ## How it works