From 1afc2d510526bf149d581d3251cb9a91618c9bda Mon Sep 17 00:00:00 2001
From: Veronica Berglyd Olsen <1619840+vkbo@users.noreply.github.com>
Date: Wed, 19 Aug 2026 19:10:31 +0200
Subject: [PATCH] Fix dropped PGS acquisition frames when there is no
 compositions frame

---
 subtle_gui/formats/pgssubs.py | 4 +++-
 subtle_gui/ocr/tesseract.py   | 2 +-
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/subtle_gui/formats/pgssubs.py b/subtle_gui/formats/pgssubs.py
index a8d9cf1..8923d67 100644
--- a/subtle_gui/formats/pgssubs.py
+++ b/subtle_gui/formats/pgssubs.py
@@ -122,8 +122,10 @@ def _readData(self, path: Path) -> None:
         frames = []
         for ds in data:
             state = ds.pcs.compState
-            if state == COMP_EPOCH:
+            if state == COMP_EPOCH or (state == COMP_ACQ and not frame):
                 # Start of a new subtitles frame.
+                # We assume that acquisition frames after closed frames are also valid subtitles.
+                # These sometimes occur, and I'm not really sure why.
                 frame = PGSFrame(len(frames), ds)
                 frames.append(frame)
             elif ds.isClearFrame() and frame:
diff --git a/subtle_gui/ocr/tesseract.py b/subtle_gui/ocr/tesseract.py
index 17bb20f..df48082 100644
--- a/subtle_gui/ocr/tesseract.py
+++ b/subtle_gui/ocr/tesseract.py
@@ -51,7 +51,7 @@
 
 RX_REPLACE = {
     "all": [
-        (re.compile(r"^(-\s)[\w]", re.UNICODE), "-"),
+        # (re.compile(r"^(-\s)[\w]", re.UNICODE), "-"),
         (re.compile(r"^(\.{2})[\s\w]", re.UNICODE), "..."),
         (re.compile(r"^(\.{3}\s)\w", re.UNICODE), "..."),
         # Wrong capitalisation in the middle of words