Add Single File Demo (Sample)

app.py

@@ -28,7 +28,7 @@ from state import (
     init_session_state,
     get_display_name, apply_speaker_renames_to_df, convert_df,
     updateMultiSelect, store_speaker_clips, register_file, analyze,
-    load_demo_single, load_demo_multi, run_analysis_loop,
+    load_demo_single, load_demo_single_sample, load_demo_multi, run_analysis_loop,
     build_all_csv_zip,
 )
 
@@ -50,6 +50,7 @@ PARQUET_DATASET_DIR = Path("parquet_dataset")
 PARQUET_DATASET_DIR.mkdir(parents=True, exist_ok=True)
 
 DEMO_PATH        = "sample.rttm"
+DEMO_SAMPLE_PATH = "sample_short.rttm"
 MULTI_DEMO_PATHS = [
     "audioSamples/media-afc-cal-afc1986022_sr01a05.rttm",
     "audioSamples/media-afc-cal-afc1986022_sr34a01.rttm",
@@ -124,7 +125,8 @@ if not isGPU:
 
 st.write(
     'If you would like to see a sample result or multiple sample results generated from '
-    'real classroom audio, select "Single File Demo" or "Multiple Files Demo" on the left sidebar.'
+    'real classroom audio, select "Single File Demo (Sample)", "Single File Demo (Full)", '
+    'or "Multiple Files Demo" on the left sidebar.'
 )
 st.markdown(
     "<p style='margin-bottom:4px;'>Keep in mind that this is a very early draft of the tool. "
@@ -177,7 +179,11 @@ file_paths_dict = st.session_state.file_paths
 
 isDemo = False
 
-if st.sidebar.button("Single File Demo"):
+if st.sidebar.button("Single File Demo (Sample)"):
+    load_demo_single_sample(DEMO_SAMPLE_PATH)
+    isDemo = True
+
+if st.sidebar.button("Single File Demo (Full)"):
     load_demo_single(DEMO_PATH)
     isDemo = True
 
@@ -185,6 +191,12 @@ if st.sidebar.button("Multiple Files Demo"):
     load_demo_multi(MULTI_DEMO_PATHS)
     isDemo = True
 
+st.sidebar.caption(
+    "Single File Demo (Sample) analyzes a 10-minute sample, while "
+    "Single File Demo (Full) analyzes the entire file. "
+    "The latter can take a while to finish."
+)
+
 # ---------------------------------------------------------------------------
 # Analyze All / Reset buttons
 # ---------------------------------------------------------------------------
@@ -240,7 +252,7 @@ try:
         raise ValueError("No file selected")
 
     st.session_state.resetResult = False
-    currPlainName = currFile.split(".")[0]
+    currPlainName = currFile.rsplit(".", 1)[0]
 
     if not (
         currFile in st.session_state.results
@@ -300,23 +312,32 @@ try:
         for sp in st.session_state.results[fn][0].labels()
     ]
     # Map display label -> raw token so ui.py can translate back after selection.
-    # Build carefully to avoid key collisions: if two speakers in the same file
-    # share a display name, append the raw ID to disambiguate both entries.
+    # Display label uses a truncated filename (first 5 chars + "...") to keep
+    # the dropdown readable when filenames are long.  The raw token always uses
+    # the full filename so the data model is never affected by this truncation.
+    # If two speakers in the same file share a display name, the raw speaker ID
+    # is appended to disambiguate both entries.
+    def _short_fn(fname):
+        """Return first 5 chars of filename stem + '...' if longer than 5."""
+        stem = fname.rsplit(".", 1)[0]   # strip extension
+        return (stem[:5] + "...") if len(stem) > 5 else stem
+
     token_display_map = {}
     for fn in st.session_state.file_names:
         if not (fn in st.session_state.results and len(st.session_state.results[fn]) == 2):
             continue
+        short = _short_fn(fn)
         # First pass: collect display labels per file to detect duplicates
-        sp_labels = {sp: f"{fn}: {get_display_name(sp, fn)}"
+        sp_labels = {sp: f"{short}: {get_display_name(sp, fn)}"
                      for sp in st.session_state.results[fn][0].labels()}
         label_counts = {}
         for disp in sp_labels.values():
             label_counts[disp] = label_counts.get(disp, 0) + 1
         # Second pass: build map, disambiguating where needed
         for sp, disp in sp_labels.items():
-            raw = f"{fn}: {sp}"
+            raw = f"{fn}: {sp}"          # raw token always uses full filename
             if label_counts[disp] > 1:
-                token_display_map[f"{fn}: {get_display_name(sp, fn)} ({sp})"] = raw
+                token_display_map[f"{short}: {get_display_name(sp, fn)} ({sp})"] = raw
             else:
                 token_display_map[disp] = raw
     display_speaker_tokens = list(token_display_map.keys())
@@ -439,7 +460,7 @@ with st.expander("(Potentially) FAQ"):
     st.write("**3. I still don't have a file to select in the dropdown! Why?**")
     st.write("Your file may be too large. We currently support approximately 1.5 hours of audio.")
     st.write("**4. I want to view my previously analyzed data. How?**")
-    st.write("Download a CSV copy from the Data tab and re-upload it later.")
+    st.write("Download a CSV copy from the Download tab and re-upload it later.")
     st.write("**5. The app is extremely slow. What is wrong?**")
     st.write("We are securing funding for permanent GPU access. Until then, CPU analysis may take a very long time.")