Add output format abstraction and diffable output formatv0.3.0

4 files changed, 36 insertions, 9 deletions

diff --git a/Justfile b/Justfile
index 3a33629..48e18e2 100644
--- a/Justfile
+++ b/Justfile
@@ -7,7 +7,7 @@ check_style:
 	black --check src
 
 build: check
-	pyproject-build
+	python3 -m build
 
 format:
 	black src
diff --git a/pyproject.toml b/pyproject.toml
index b175e02..b660215 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [project]
 name = "srtfilter"
-version = "0.2.1"
+version = "0.3.0"
 authors = [
     { name="Arjun Satarkar", email="me@arjunsatarkar.net" },
 ]
diff --git a/src/srtfilter/parse.py b/src/srtfilter/parse.py
index cbfb5ea..b53bb37 100644
--- a/src/srtfilter/parse.py
+++ b/src/srtfilter/parse.py
@@ -1,8 +1,15 @@
 from __future__ import annotations
 import dataclasses
+import enum
 import re
 
 
+class OutputFormat(enum.Enum):
+    none = enum.auto()
+    srt = enum.auto()
+    diffable_srt = enum.auto()
+
+
 class SRT:
     # Can't use the one from Timecode because that has named groups, which can't be duplicates
     timecode_pattern_str = r"\d\d:\d\d:\d\d,\d\d\d"
@@ -40,11 +47,24 @@ class SRT:
         return srt
 
     def __str__(self):
+        return self.to_output_format(OutputFormat.srt)
+
+    def to_output_format(self, format: OutputFormat) -> str:
         result = ""
-        for counter, event in enumerate(self.events, 1):
-            result += f"{counter}\n"
-            result += f"{event.start} --> {event.end}\n"
-            result += f"{event.content}\n"
+        match format:
+            case OutputFormat.none:
+                pass
+            case OutputFormat.srt:
+                for counter, event in enumerate(self.events, 1):
+                    result += f"{counter}\n"
+                    result += f"{event.start} --> {event.end}\n"
+                    result += f"{event.content}\n"
+            case OutputFormat.diffable_srt:
+                # Not an actual subtitle format, just drops the counter to
+                # make it easier to diff actual changes (in timing/content).
+                for event in self.events:
+                    result += f"{event.start} --> {event.end}\n"
+                    result += f"{event.content}\n"
         return result
 
 
diff --git a/src/srtfilter/srtfilter_cli.py b/src/srtfilter/srtfilter_cli.py
index bc44409..30aa942 100755
--- a/src/srtfilter/srtfilter_cli.py
+++ b/src/srtfilter/srtfilter_cli.py
@@ -8,8 +8,15 @@ from .filters import rebreak_lines
 @click.command()
 @click.argument("in_file_path")
 @click.option("--filter", "filter_arg", default="")
-def main(in_file_path: str, filter_arg: str):
-    with open(in_file_path) as f:
+@click.option(
+    "--out-format",
+    type=click.Choice([v.name for v in parse.OutputFormat]),
+    default="srt",
+)
+def main(in_file_path: str, filter_arg: str, out_format: str):
+    # Encoding utf-8-sig to handle BOM, which Aegisub seems to generate
+    # when exporting to SRT. This works fine even if there is no BOM.
+    with open(in_file_path, encoding="utf-8-sig") as f:
         text = f.read()
     srt = parse.SRT.from_str(text)
 
@@ -21,7 +28,7 @@ def main(in_file_path: str, filter_arg: str):
                 raise InvalidFilterError(unknown)
         srt.events = [filter_module.filter(event) for event in srt.events]
 
-    sys.stdout.write(str(srt))
+    sys.stdout.write(srt.to_output_format(parse.OutputFormat[out_format]))
 
 
 class InvalidFilterError(Exception):