From 9cc4d1c526100a4a20f6c8f0ff842459c1611094 Mon Sep 17 00:00:00 2001
From: Owen-Cochell <owencochell@gmail.com>
Date: Wed, 12 Apr 2023 17:22:50 -0400
Subject: [PATCH] Added time interval merging

---
 static/js/transcript-editing.js |  2 +-
 transcription/views.py          | 73 +++++++++++++++++++++++++++++++--
 2 files changed, 71 insertions(+), 4 deletions(-)

diff --git a/static/js/transcript-editing.js b/static/js/transcript-editing.js
index b25b9d2..b78e6e4 100755
--- a/static/js/transcript-editing.js
+++ b/static/js/transcript-editing.js
@@ -88,7 +88,7 @@ function deleteTranscript() {
     // if any word has been selected, then update the video
         selectedWordsArr.forEach(element => {
             console.log(element);
-            timestamps["timestamps"].push([element.getAttribute("data-start"), element.getAttribute("data-stop")]);
+            timestamps["timestamps"].push([parseFloat(element.getAttribute("data-start")), parseFloat(element.getAttribute("data-stop"))]);
             element.remove();
         });
 
diff --git a/transcription/views.py b/transcription/views.py
index 7008838..930b4a8 100644
--- a/transcription/views.py
+++ b/transcription/views.py
@@ -10,12 +10,67 @@
 import whisper_timestamped as whisper
 import moviepy.editor
 
+SMALL = 0.1  # Number of seconds for two cuts to count as one
 
 # Load a model:
 
 model = whisper.load_model("base")
 
 
+def merge_times(times):
+    """
+    Merges a list of times.
+
+    The idea is to convert two cuts with a small
+    time diff into one cut to avoid awkward cuts.
+
+    :param times: List of times
+    :type times: List[Tuple(int, int)]
+    """
+
+    final = []
+    last = [0, 0]
+    index = -1
+
+    # Iterate over all times:
+
+    while index < len(times) - 1:
+
+        # Determine if we need to merge
+
+        if times[index+1][0] - last[1] < SMALL:
+
+            # Set last:
+
+            last = [last[0], times[index+1][1]]
+
+        # Otherwise, just set last:
+
+        else:
+
+            # Determine if we should add to final:
+
+            if last[0] != last[1]:
+
+                # Add value to final:
+
+                final.append(last)
+
+            # Set last:
+
+            last = times[index+1]
+
+        # Increment index:
+
+        index += 1
+
+    # Finally, add last to list:
+
+    final.append(last)
+
+    return final
+
+
 def index(request):
     return render(request, 'Frontend/Home.html')
 
@@ -174,7 +229,19 @@ def cut_files(request):
 
         # Decode the cuts:
 
-        cuts = json.loads(request.POST['cuts'])
+        cuts = json.loads(request.POST['cuts'])['timestamps']
+
+        # Sort by start value:
+
+        cuts.sort(key=lambda x: x[0])
+
+        print("Original Cut: {}".format(cuts))
+
+        if len(cuts) >= 2:
+
+            cuts = merge_times(cuts)
+
+        print(cuts)
 
         # Create video model and blank file:
 
@@ -190,8 +257,8 @@ def cut_files(request):
 
         # Iterate over each cut:
 
-        for cut in cuts['timestamps']:
-            cut = [eval(i) for i in cut]
+        for cut in cuts:
+
             print(f"Current timestamp: {cut}")
 
             clip = clip.cutout(cut[0] - total_removed, cut[1] - total_removed)