Merge pull request #6 from Goekdeniz-Guelmez/adding-three-to-five-speacker-formats

v1.1.0

Author: Goekdeniz-Guelmez

README.md

@@ -2,7 +2,7 @@
 
 ![logo](logo.jpeg)
 
-A local AI-powered tool that converts PDF documents into engaging podcasts, using local LLMs and TTS models.
+A local AI-powered tool that converts PDF documents into engaging audio's such as podcasts, using local LLMs and TTS models.
 
 ## Features
 
@@ -68,7 +68,10 @@ You can use the default configuration or create a custom JSON config file with t
 
 ```json
 {
-    "Co-Host-Speaker-Voice": "af_sky+af_bella",
+    "Co-Host-Speaker-1-Voice": "af_sky+af_bella",
+    "Co-Host-Speaker-2-Voice": "af_echo",
+    "Co-Host-Speaker-3-Voice": "af_nova",
+    "Co-Host-Speaker-4-Voice": "af_shimmer",
     "Host-Speaker-Voice": "af_alloy",
 
     "Small-Text-Model": {
@@ -224,10 +227,9 @@ python -m local_notebooklm.start --pdf PATH_TO_PDF [options]
 | `--length` | Content length (short, medium, long, very-long) | medium |
 | `--style` | Content style (normal, casual, formal, technical, academic, friendly, gen-z, funny) | normal |
 | `--preference` | Additional focus preferences or instructions | None |
+| `--language` | Language the audio should be in | english |
 | `--output-dir` | Directory to store output files | ./output |
 
-Local-NotebookLM currently does NOT support multible languages other then english, you can try working around it by adding a text in the preferences saying what language the audio should be, also be sure the TTS model supports your desired language.
-
 #### Format Types
 
 Local-NotebookLM now supports both single-speaker and two-speaker formats:
@@ -251,6 +253,10 @@ Local-NotebookLM now supports both single-speaker and two-speaker formats:
 - q-and-a
 - meeting
 
+**Multi-Speaker Formats:**
+- panel-discussion (3, 4, or 5 speakers)
+- debate (3, 4, or 5 speakers)
+
 #### Example Commands
 
 Basic usage:
@@ -270,7 +276,7 @@ python -m local_notebooklm.start --pdf documents/research_paper.pdf --preference
 
 Using custom config:
 ```bash
-python -m local_notebooklm.start --pdf documents/research_paper.pdf --config custom_config.json --output-dir ./my_podcast
+python -m local_notebooklm.start --pdf documents/research_paper.pdf --config custom_config.json --output-dir ./my_podcast --language german
 ```
 
 ### Programmatic API
@@ -287,7 +293,8 @@ success, result = podcast_processor(
     length="long",
     style="professional",
     preference="Focus on the key technical aspects",
-    output_dir="./test_output"
+    output_dir="./test_output",
+    language="english"
 )
 
 if success:
@@ -441,6 +448,13 @@ flowchart TD
     class pdf,file1,file2,file3,fileAudio data
 ```
 
+## Multiple Language Support
+
+Local-NotebookLM now supports multiple languages. You can specify the language when using the programmatic API or through the command line.
+
+**Important Note:** When using a non-English language, ensure that both your selected LLM and TTS models support the desired language. Language support varies significantly between different models and providers. For optimal results, verify that your chosen models have strong capabilities in your target language before processing.
+
+
 ## Output Files
 
 The pipeline generates the following files:

local_notebooklm/processor.py

@@ -14,7 +14,8 @@ def podcast_processor(
     style="normal",
     preference="nothing",
     output_dir="./output",
-    skip_to=None
+    skip_to: int = None,
+    language: str = "english"
 ):
     # Load config
     if config_path:
@@ -116,7 +117,8 @@ def podcast_processor(
                 input_file=transcript_file,
                 output_dir=str(output_dirs["step3"]),
                 format_type=format_type,
-                system_prompt=system_prompts["step3"]
+                system_prompt=system_prompts["step3"],
+                language=language
             )
         else:
             print("Skipping Step 3, assuming files exist in output directory...")

local_notebooklm/steps/helpers.py

@@ -13,24 +13,51 @@
     "lecture", "tutorial", "q-and-a",
     "news-report", "executive-brief", "meeting", "analysis"
 ]
+
+
 SingleSpeakerFormats = Literal[
     "summary", "narration", "storytelling", "explainer",
     "lecture", "tutorial", "news-report", "executive-brief", "analysis"
 ]
-TwoSpeakerFormats = Literal[
-    "podcast", "interview", "panel-discussion",
-    "debate", "q-and-a", "meeting"
-]
 SINGLE_SPEAKER_FORMATS = {
     "summary", "narration", "storytelling", "explainer",
     "lecture", "tutorial", "news-report", "executive-brief", "analysis"
 }
 
+TwoSpeakerFormats = Literal[
+    "podcast", "interview", "panel-discussion",
+    "debate", "q-and-a", "meeting"
+]
 TWO_SPEAKER_FORMATS = {
     "podcast", "interview", "panel-discussion",
     "debate", "q-and-a", "meeting"
 }
+
+ThreeSpeakerFormats = Literal[
+    "three-people-podcast", "three-people-panel-discussion", "three-people-debate"
+]
+THREE_SPEAKER_FORMATS = {
+    "three-people-podcast", "three-people-panel-discussion", "three-people-debate"
+}
+
+FourSpeakerFormats = Literal[
+    "four-people-podcast", "four-people-panel-discussion", "four-people-debate"
+]
+FOUR_SPEAKER_FORMATS = {
+    "four-people-podcast", "four-people-panel-discussion", "four-people-debate"
+}
+
+FiveSpeakerFormats = Literal[
+    "five-people-podcast", "five-people-panel-discussion", "five-people-debate"
+]
+FIVE_SPEAKER_FORMATS = {
+    "five-people-podcast", "five-people-panel-discussion", "five-people-debate"
+}
+
+
 LengthType = Literal["short", "medium", "long", "very-long"]
+
+
 StyleType = Literal["normal", "friendly", "professional", "academic", "casual", "technical", "gen-z", "funny"]