dwani 0.1.1__tar.gz → 0.1.3__tar.gz

dwani-0.1.3/PKG-INFO

@@ -0,0 +1,70 @@
+Metadata-Version: 2.4
+Name: dwani
+Version: 0.1.3
+Summary: Multimodal AI server for Indian languages (speech, vision, LLMs, TTS, ASR, etc.)
+Author-email: sachin <python@dwani.ai>
+License: MIT License
+        
+        Copyright (c) 2025 Sachin Shetty
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+        
+Project-URL: Homepage, https://github.com/dwani-ai/dwani-python
+Project-URL: Source, https://github.com/dwani-ai/dwani-python
+Project-URL: Issues, https://github.com/dwani-ai/dwani-python/issues
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: requests>=2.25.0
+Dynamic: license-file
+
+# dwani.ai - python library
+
+
+```bash
+pip install dwani
+```
+
+
+
+```python
+import dwani
+import os
+
+dwani.api_key = os.getenv("DWANI_API_KEY")
+
+dwani.api_base = os.getenv("DWANI_API_BASE")
+
+resp = dwani.Chat.create("Hello!", "eng_Latn", "kan_Knda")
+print(resp)
+```
+
+
+<!-- 
+## local development
+pip install -e .
+
+
+pip install twine build
+rm -rf dist/
+python -m build
+
+python -m twine upload dist/*
+
+-->

dwani-0.1.3/README.md

@@ -0,0 +1,34 @@
+# dwani.ai - python library
+
+
+```bash
+pip install dwani
+```
+
+
+
+```python
+import dwani
+import os
+
+dwani.api_key = os.getenv("DWANI_API_KEY")
+
+dwani.api_base = os.getenv("DWANI_API_BASE")
+
+resp = dwani.Chat.create("Hello!", "eng_Latn", "kan_Knda")
+print(resp)
+```
+
+
+<!-- 
+## local development
+pip install -e .
+
+
+pip install twine build
+rm -rf dist/
+python -m build
+
+python -m twine upload dist/*
+
+-->

{dwani-0.1.1/src → dwani-0.1.3/dwani}/asr.py

@@ -1,5 +1,5 @@
 from .exceptions import DhwaniAPIError
-
+import requests
 def asr_transcribe(client, file_path, language):
     with open(file_path, "rb") as f:
         files = {"file": f}
@@ -18,6 +18,8 @@ class ASR:
         from . import _get_client
         return _get_client().transcribe(*args, **kwargs)
 
+
+'''
 from .docs import Documents
 
 class documents:
@@ -32,3 +34,4 @@ class documents:
     @staticmethod
     def summarize(file_path, language=None):
         return _get_client().document_summarize(file_path, language)
+'''

{dwani-0.1.1/src → dwani-0.1.3/dwani}/audio.py

@@ -1,5 +1,5 @@
 from .exceptions import DhwaniAPIError
-
+import requests
 def audio_speech(client, input, voice, model, response_format="mp3", output_file=None):
     data = {
         "input": input,

dwani-0.1.3/dwani/chat.py

@@ -0,0 +1,25 @@
+from .exceptions import DhwaniAPIError
+import requests
+
+def chat_create(client, prompt, src_lang, tgt_lang, **kwargs):
+    url = f"{client.api_base}/v1/indic_chat"
+    payload = {
+        "prompt": prompt,
+        "src_lang": src_lang,
+        "tgt_lang": tgt_lang
+    }
+    payload.update(kwargs)
+    resp = requests.post(
+        url,
+        headers={**client._headers(), "Content-Type": "application/json"},
+        json=payload
+    )
+    if resp.status_code != 200:
+        raise DhwaniAPIError(resp)
+    return resp.json()
+
+class Chat:
+    @staticmethod
+    def create(prompt, src_lang, tgt_lang, **kwargs):
+        from . import _get_client
+        return _get_client().chat(prompt, src_lang, tgt_lang, **kwargs)

{dwani-0.1.1/src → dwani-0.1.3/dwani}/client.py

@@ -4,17 +4,18 @@ from .exceptions import DhwaniAPIError
 
 class DhwaniClient:
     def __init__(self, api_key=None, api_base=None):
-        self.api_key = api_key or os.getenv("DHWANI_API_KEY")
-        self.api_base = api_base or os.getenv("DHWANI_API_BASE", "http://localhost:7860")
+        self.api_key = api_key or os.getenv("DWANI_API_KEY")
+        self.api_base = api_base or os.getenv("DWANI_API_BASE", "http://localhost:7860")
         if not self.api_key:
             raise ValueError("DHWANI_API_KEY not set")
 
     def _headers(self):
         return {"X-API-Key": self.api_key}
 
-    def chat(self, prompt, **kwargs):
+    def chat(self, prompt, src_lang, tgt_lang, **kwargs):
         from .chat import chat_create
-        return chat_create(self, prompt, **kwargs)
+        return chat_create(self, prompt, src_lang, tgt_lang, **kwargs)
+
 
     def speech(self, *args, **kwargs):
         from .audio import audio_speech

{dwani-0.1.1/src → dwani-0.1.3/dwani}/vision.py

@@ -1,5 +1,5 @@
 from .exceptions import DhwaniAPIError
-
+import requests
 def vision_caption(client, file_path, length="short"):
     with open(file_path, "rb") as f:
         files = {"file": f}

dwani-0.1.3/dwani.egg-info/PKG-INFO

@@ -0,0 +1,70 @@
+Metadata-Version: 2.4
+Name: dwani
+Version: 0.1.3
+Summary: Multimodal AI server for Indian languages (speech, vision, LLMs, TTS, ASR, etc.)
+Author-email: sachin <python@dwani.ai>
+License: MIT License
+        
+        Copyright (c) 2025 Sachin Shetty
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+        
+Project-URL: Homepage, https://github.com/dwani-ai/dwani-python
+Project-URL: Source, https://github.com/dwani-ai/dwani-python
+Project-URL: Issues, https://github.com/dwani-ai/dwani-python/issues
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: requests>=2.25.0
+Dynamic: license-file
+
+# dwani.ai - python library
+
+
+```bash
+pip install dwani
+```
+
+
+
+```python
+import dwani
+import os
+
+dwani.api_key = os.getenv("DWANI_API_KEY")
+
+dwani.api_base = os.getenv("DWANI_API_BASE")
+
+resp = dwani.Chat.create("Hello!", "eng_Latn", "kan_Knda")
+print(resp)
+```
+
+
+<!-- 
+## local development
+pip install -e .
+
+
+pip install twine build
+rm -rf dist/
+python -m build
+
+python -m twine upload dist/*
+
+-->

dwani-0.1.3/dwani.egg-info/SOURCES.txt

@@ -0,0 +1,16 @@
+LICENSE
+README.md
+pyproject.toml
+dwani/__init__.py
+dwani/asr.py
+dwani/audio.py
+dwani/chat.py
+dwani/client.py
+dwani/docs.py
+dwani/exceptions.py
+dwani/vision.py
+dwani.egg-info/PKG-INFO
+dwani.egg-info/SOURCES.txt
+dwani.egg-info/dependency_links.txt
+dwani.egg-info/requires.txt
+dwani.egg-info/top_level.txt

dwani-0.1.3/dwani.egg-info/requires.txt

@@ -0,0 +1 @@
+requests>=2.25.0

dwani-0.1.3/dwani.egg-info/top_level.txt

@@ -0,0 +1 @@
+dwani

{dwani-0.1.1 → dwani-0.1.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "dwani"
-version = "0.1.1"
+version = "0.1.3"
 description = "Multimodal AI server for Indian languages (speech, vision, LLMs, TTS, ASR, etc.)"
 authors = [
     { name="sachin", email="python@dwani.ai" }
@@ -14,15 +14,10 @@ license = { file = "LICENSE" }
 requires-python = ">=3.10"
 
 dependencies = [
-    "fastapi>=0.95.0",
-    "uvicorn[standard]>=0.22.0",
-    "pydantic>=2.0.0",
     "requests>=2.25.0",
-    "python-multipart>=0.0.5",
-    "pydantic-settings>=2.0.0"
 ]
 
 [project.urls]
-Homepage = "https://github.com/dwani-ai/dwani-server"
-Source = "https://github.com/dwani-ai/dwani-server"
-Issues = "https://github.com/dwani-ai/dwani-server/issues"
+Homepage = "https://github.com/dwani-ai/dwani-python"
+Source = "https://github.com/dwani-ai/dwani-python"
+Issues = "https://github.com/dwani-ai/dwani-python/issues"

dwani-0.1.1/PKG-INFO

@@ -1,193 +0,0 @@
-Metadata-Version: 2.4
-Name: dwani
-Version: 0.1.1
-Summary: Multimodal AI server for Indian languages (speech, vision, LLMs, TTS, ASR, etc.)
-Author-email: sachin <python@dwani.ai>
-License: MIT License
-        
-        Copyright (c) 2025 Sachin Shetty
-        
-        Permission is hereby granted, free of charge, to any person obtaining a copy
-        of this software and associated documentation files (the "Software"), to deal
-        in the Software without restriction, including without limitation the rights
-        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-        copies of the Software, and to permit persons to whom the Software is
-        furnished to do so, subject to the following conditions:
-        
-        The above copyright notice and this permission notice shall be included in all
-        copies or substantial portions of the Software.
-        
-        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-        SOFTWARE.
-        
-Project-URL: Homepage, https://github.com/dwani-ai/dwani-server
-Project-URL: Source, https://github.com/dwani-ai/dwani-server
-Project-URL: Issues, https://github.com/dwani-ai/dwani-server/issues
-Requires-Python: >=3.10
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: fastapi>=0.95.0
-Requires-Dist: uvicorn[standard]>=0.22.0
-Requires-Dist: pydantic>=2.0.0
-Requires-Dist: requests>=2.25.0
-Requires-Dist: python-multipart>=0.0.5
-Requires-Dist: pydantic-settings>=2.0.0
-Dynamic: license-file
-
-# Dhwani Server
-
-Dhwani API is a FastAPI-based application providing AI-powered services for Indian languages, including text-to-speech (TTS), language model (LLM) chat, vision-language model (VLM) capabilities, and automatic speech recognition (ASR). It supports lazy loading of models for fast startup and includes endpoints for various tasks.
-
-## Features
-- **Text-to-Speech (TTS)**: Generate audio from text in Indian languages using Parler TTS.
-- **Chat**: Process Kannada prompts and respond in Kannada via translation and LLM.
-- **Vision-Language Model (VLM)**: Caption images, answer visual queries, detect, and point objects.
-- **Automatic Speech Recognition (ASR)**: Transcribe audio files in multiple Indian languages.
-- **Lazy Loading**: Models load on-demand or via an explicit endpoint for fast startup.
-
-## Prerequisites
-- **System Requirements - User **:
-    - **Python**: 3.10
-    - Ubuntu 22.04
-    - git 
-    - vscode
-- **System Requirements - Server **:
-  - Ubuntu with sufficient RAM (16GB+ recommended for models).
-  - Optional: NVIDIA GPU with CUDA support for faster inference.
-- **FFmpeg**: Required for audio processing (ASR).
-
-- Server Setup
-```bash
-export HF_HOME=/home/ubuntu/data-dhwani-models
-export HF_TOKEN='YOur-HF-token'
-python src/server/main.py --host 0.0.0.0 --port 7860 --config config_two
-```
-## Installation
-
-1. **Clone the Repository**:
-   ```bash
-   git clone https://github.com/slabstech/dhwani-server
-   cd dhwani-server
-   ```
-
-2. Install Libraries:
-    - On Ubuntu: ```sudo apt-get install ffmpeg build-essential```
-
-3. Set Up Virtual Environment:
-    ```bash
-    python -m venv venv
-    source venv/bin/activate 
-    ```
-4. Install Dependencies:
-    ```bash
-    sudo apt-get install -y ffmpeg build-essential
-    curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --no-modify-path --profile minimal
-    . "$HOME/.cargo/env"
-    export CC=/usr/bin/gcc
-    export ENV CXX=/usr/bin/g++
-    ```
-    ```bash
-    pip install --no-cache-dir --upgrade pip setuptools psutil setuptools-rust torch==2.6.0
-    pip install --no-cache-dir flash-attn --no-build-isolation
-    ```
-
-    ```bash
-    pip install -r requirements.txt
-    ```
-    
-4. Set Environment Variable:
-    Create a .env file in the root directory and add your API key:
-    plaintext
-    ```bash
-    API_KEY=your_secret_key
-    ```
-
-5. Running the Server
-    - Start the Server:
-    ```bash
-        python src/server/main.py --host 0.0.0.0 --port 7860 --config config_two
-    ```
-
-    - The server starts with models loaded on start
-    - Access the interactive API docs at http://localhost:7860/docs.
-    
-    - (Optional) Load All Models:
-    Preload all models (LLM, Translation, TTS, VLM, ASR) with:
-    - 
-    ```bash
-    curl -X POST "http://localhost:7860/load_all_models" -H "X-API-Key: your_secret_key"
-    ```
-
-- Usage
-    - Endpoints
-        - All endpoints require the X-API-Key header with the value from your .env file.
-
-    -    Health Check: GET /health
-    ```bash
-        curl "http://localhost:7860/health"
-    ```
-    - Response: 
-        ```bash 
-        {"status": "healthy", "model": "Qwen/Qwen2.5-3B-Instruct"}
-        ```
-    
-    - Text-to-Speech: POST /v1/audio/speech
-    ```  bash
-    curl -X POST "http://localhost:7860/v1/audio/speech" -H "X-API-Key: your_secret_key" -H "Content-Type: application/json" -d '{"input": "ನಮಸ್ಕಾರ", "voice": "Female voice", "model": "ai4bharat/indic-parler-tts", "response_format": "mp3"}' --output speech.mp3
-    ```
-    - Chat: POST /chat
-    ``` bash
-        curl -X POST "http://localhost:7860/chat" -H "X-API-Key: your_secret_key" -H "Content-Type: application/json" -d '{"prompt": "ನೀವು ಹೇಗಿದ್ದೀರಿ?"}'
-    ```
-
-    - Response: 
-            ```{"response": "<Kannada response>"}```
-    - Image Captioning: POST /caption/
-    ```bash
-        curl -X POST "http://localhost:7860/caption/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "length=short"
-    ```
-    - Response:``` {"caption": "<short caption>"}```
-    - Visual Query: POST /visual_query/
-    ```bash
-        curl -X POST "http://localhost:7860/visual_query/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "query=What is this?"
-    ```
-    - Response: ```{"answer": "<answer>"}```
-    - Object Detection: POST /detect/
-    ```bash
-        curl -X POST "http://localhost:7860/detect/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "object_type=face"
-    ```
-    - Response: ```{"objects": [<list of detected objects>]}```
-    - Object Pointing: POST /point/
-    ```bash
-
-        curl -X POST "http://localhost:7860/point/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "object_type=person"
-    ```
-    - Response: ```{"points": [<list of points>]}```
-    - Transcription: POST /transcribe/
-    ```bash
-    curl -X POST "http://localhost:7860/transcribe/?language=kannada" -H "X-API-Key: your_secret_key" -F "file=@audio.wav"
-    ```
-    - Response: ```{"text": "<transcribed text>"}```
-    - Batch Transcription: POST /transcribe_batch/
-    ```bash
-    curl -X POST "http://localhost:7860/transcribe_batch/?language=kannada" -H "X-API-Key: your_secret_key" -F "files=@audio1.wav" -F "files=@audio2.mp3"
-    ```
-    - Response: ```{"transcriptions": ["<text1>", "<text2>"]}```
-
-- Notes
-    - Lazy Loading: Models load on first use or via /load_all_models. Expect a delay on the first request for each model type.
-    Supported Languages: ASR supports multiple Indian languages (e.g., kannada, hindi, tamil); see models/asr.py for the full list.
-    Logs: Check dhwani_api.log for detailed logs (rotated at 10MB, 5 backups).
-    Performance: Use a GPU with flash-attn installed for faster TTS and ASR inference.
-
-- Troubleshooting
-
-    - Module Errors: Ensure all dependencies are installed. Re-run pip install if needed.
-    FFmpeg Not Found: Install FFmpeg and ensure it’s in your PATH.
-    Permission Denied: Run with sudo if accessing restricted ports (e.g., < 1024).
-

dwani-0.1.1/README.md

@@ -1,152 +0,0 @@
-# Dhwani Server
-
-Dhwani API is a FastAPI-based application providing AI-powered services for Indian languages, including text-to-speech (TTS), language model (LLM) chat, vision-language model (VLM) capabilities, and automatic speech recognition (ASR). It supports lazy loading of models for fast startup and includes endpoints for various tasks.
-
-## Features
-- **Text-to-Speech (TTS)**: Generate audio from text in Indian languages using Parler TTS.
-- **Chat**: Process Kannada prompts and respond in Kannada via translation and LLM.
-- **Vision-Language Model (VLM)**: Caption images, answer visual queries, detect, and point objects.
-- **Automatic Speech Recognition (ASR)**: Transcribe audio files in multiple Indian languages.
-- **Lazy Loading**: Models load on-demand or via an explicit endpoint for fast startup.
-
-## Prerequisites
-- **System Requirements - User **:
-    - **Python**: 3.10
-    - Ubuntu 22.04
-    - git 
-    - vscode
-- **System Requirements - Server **:
-  - Ubuntu with sufficient RAM (16GB+ recommended for models).
-  - Optional: NVIDIA GPU with CUDA support for faster inference.
-- **FFmpeg**: Required for audio processing (ASR).
-
-- Server Setup
-```bash
-export HF_HOME=/home/ubuntu/data-dhwani-models
-export HF_TOKEN='YOur-HF-token'
-python src/server/main.py --host 0.0.0.0 --port 7860 --config config_two
-```
-## Installation
-
-1. **Clone the Repository**:
-   ```bash
-   git clone https://github.com/slabstech/dhwani-server
-   cd dhwani-server
-   ```
-
-2. Install Libraries:
-    - On Ubuntu: ```sudo apt-get install ffmpeg build-essential```
-
-3. Set Up Virtual Environment:
-    ```bash
-    python -m venv venv
-    source venv/bin/activate 
-    ```
-4. Install Dependencies:
-    ```bash
-    sudo apt-get install -y ffmpeg build-essential
-    curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --no-modify-path --profile minimal
-    . "$HOME/.cargo/env"
-    export CC=/usr/bin/gcc
-    export ENV CXX=/usr/bin/g++
-    ```
-    ```bash
-    pip install --no-cache-dir --upgrade pip setuptools psutil setuptools-rust torch==2.6.0
-    pip install --no-cache-dir flash-attn --no-build-isolation
-    ```
-
-    ```bash
-    pip install -r requirements.txt
-    ```
-    
-4. Set Environment Variable:
-    Create a .env file in the root directory and add your API key:
-    plaintext
-    ```bash
-    API_KEY=your_secret_key
-    ```
-
-5. Running the Server
-    - Start the Server:
-    ```bash
-        python src/server/main.py --host 0.0.0.0 --port 7860 --config config_two
-    ```
-
-    - The server starts with models loaded on start
-    - Access the interactive API docs at http://localhost:7860/docs.
-    
-    - (Optional) Load All Models:
-    Preload all models (LLM, Translation, TTS, VLM, ASR) with:
-    - 
-    ```bash
-    curl -X POST "http://localhost:7860/load_all_models" -H "X-API-Key: your_secret_key"
-    ```
-
-- Usage
-    - Endpoints
-        - All endpoints require the X-API-Key header with the value from your .env file.
-
-    -    Health Check: GET /health
-    ```bash
-        curl "http://localhost:7860/health"
-    ```
-    - Response: 
-        ```bash 
-        {"status": "healthy", "model": "Qwen/Qwen2.5-3B-Instruct"}
-        ```
-    
-    - Text-to-Speech: POST /v1/audio/speech
-    ```  bash
-    curl -X POST "http://localhost:7860/v1/audio/speech" -H "X-API-Key: your_secret_key" -H "Content-Type: application/json" -d '{"input": "ನಮಸ್ಕಾರ", "voice": "Female voice", "model": "ai4bharat/indic-parler-tts", "response_format": "mp3"}' --output speech.mp3
-    ```
-    - Chat: POST /chat
-    ``` bash
-        curl -X POST "http://localhost:7860/chat" -H "X-API-Key: your_secret_key" -H "Content-Type: application/json" -d '{"prompt": "ನೀವು ಹೇಗಿದ್ದೀರಿ?"}'
-    ```
-
-    - Response: 
-            ```{"response": "<Kannada response>"}```
-    - Image Captioning: POST /caption/
-    ```bash
-        curl -X POST "http://localhost:7860/caption/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "length=short"
-    ```
-    - Response:``` {"caption": "<short caption>"}```
-    - Visual Query: POST /visual_query/
-    ```bash
-        curl -X POST "http://localhost:7860/visual_query/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "query=What is this?"
-    ```
-    - Response: ```{"answer": "<answer>"}```
-    - Object Detection: POST /detect/
-    ```bash
-        curl -X POST "http://localhost:7860/detect/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "object_type=face"
-    ```
-    - Response: ```{"objects": [<list of detected objects>]}```
-    - Object Pointing: POST /point/
-    ```bash
-
-        curl -X POST "http://localhost:7860/point/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "object_type=person"
-    ```
-    - Response: ```{"points": [<list of points>]}```
-    - Transcription: POST /transcribe/
-    ```bash
-    curl -X POST "http://localhost:7860/transcribe/?language=kannada" -H "X-API-Key: your_secret_key" -F "file=@audio.wav"
-    ```
-    - Response: ```{"text": "<transcribed text>"}```
-    - Batch Transcription: POST /transcribe_batch/
-    ```bash
-    curl -X POST "http://localhost:7860/transcribe_batch/?language=kannada" -H "X-API-Key: your_secret_key" -F "files=@audio1.wav" -F "files=@audio2.mp3"
-    ```
-    - Response: ```{"transcriptions": ["<text1>", "<text2>"]}```
-
-- Notes
-    - Lazy Loading: Models load on first use or via /load_all_models. Expect a delay on the first request for each model type.
-    Supported Languages: ASR supports multiple Indian languages (e.g., kannada, hindi, tamil); see models/asr.py for the full list.
-    Logs: Check dhwani_api.log for detailed logs (rotated at 10MB, 5 backups).
-    Performance: Use a GPU with flash-attn installed for faster TTS and ASR inference.
-
-- Troubleshooting
-
-    - Module Errors: Ensure all dependencies are installed. Re-run pip install if needed.
-    FFmpeg Not Found: Install FFmpeg and ensure it’s in your PATH.
-    Permission Denied: Run with sudo if accessing restricted ports (e.g., < 1024).
-

dwani-0.1.1/src/chat.py

@@ -1,17 +0,0 @@
-from .exceptions import DhwaniAPIError
-
-def chat_create(client, prompt, **kwargs):
-    resp = requests.post(
-        f"{client.api_base}/chat",
-        headers={**client._headers(), "Content-Type": "application/json"},
-        json={"prompt": prompt, **kwargs}
-    )
-    if resp.status_code != 200:
-        raise DhwaniAPIError(resp)
-    return resp.json()
-
-class Chat:
-    @staticmethod
-    def create(prompt, **kwargs):
-        from . import _get_client
-        return _get_client().chat(prompt, **kwargs)

dwani-0.1.1/src/dwani.egg-info/PKG-INFO

@@ -1,193 +0,0 @@
-Metadata-Version: 2.4
-Name: dwani
-Version: 0.1.1
-Summary: Multimodal AI server for Indian languages (speech, vision, LLMs, TTS, ASR, etc.)
-Author-email: sachin <python@dwani.ai>
-License: MIT License
-        
-        Copyright (c) 2025 Sachin Shetty
-        
-        Permission is hereby granted, free of charge, to any person obtaining a copy
-        of this software and associated documentation files (the "Software"), to deal
-        in the Software without restriction, including without limitation the rights
-        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-        copies of the Software, and to permit persons to whom the Software is
-        furnished to do so, subject to the following conditions:
-        
-        The above copyright notice and this permission notice shall be included in all
-        copies or substantial portions of the Software.
-        
-        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-        SOFTWARE.
-        
-Project-URL: Homepage, https://github.com/dwani-ai/dwani-server
-Project-URL: Source, https://github.com/dwani-ai/dwani-server
-Project-URL: Issues, https://github.com/dwani-ai/dwani-server/issues
-Requires-Python: >=3.10
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: fastapi>=0.95.0
-Requires-Dist: uvicorn[standard]>=0.22.0
-Requires-Dist: pydantic>=2.0.0
-Requires-Dist: requests>=2.25.0
-Requires-Dist: python-multipart>=0.0.5
-Requires-Dist: pydantic-settings>=2.0.0
-Dynamic: license-file
-
-# Dhwani Server
-
-Dhwani API is a FastAPI-based application providing AI-powered services for Indian languages, including text-to-speech (TTS), language model (LLM) chat, vision-language model (VLM) capabilities, and automatic speech recognition (ASR). It supports lazy loading of models for fast startup and includes endpoints for various tasks.
-
-## Features
-- **Text-to-Speech (TTS)**: Generate audio from text in Indian languages using Parler TTS.
-- **Chat**: Process Kannada prompts and respond in Kannada via translation and LLM.
-- **Vision-Language Model (VLM)**: Caption images, answer visual queries, detect, and point objects.
-- **Automatic Speech Recognition (ASR)**: Transcribe audio files in multiple Indian languages.
-- **Lazy Loading**: Models load on-demand or via an explicit endpoint for fast startup.
-
-## Prerequisites
-- **System Requirements - User **:
-    - **Python**: 3.10
-    - Ubuntu 22.04
-    - git 
-    - vscode
-- **System Requirements - Server **:
-  - Ubuntu with sufficient RAM (16GB+ recommended for models).
-  - Optional: NVIDIA GPU with CUDA support for faster inference.
-- **FFmpeg**: Required for audio processing (ASR).
-
-- Server Setup
-```bash
-export HF_HOME=/home/ubuntu/data-dhwani-models
-export HF_TOKEN='YOur-HF-token'
-python src/server/main.py --host 0.0.0.0 --port 7860 --config config_two
-```
-## Installation
-
-1. **Clone the Repository**:
-   ```bash
-   git clone https://github.com/slabstech/dhwani-server
-   cd dhwani-server
-   ```
-
-2. Install Libraries:
-    - On Ubuntu: ```sudo apt-get install ffmpeg build-essential```
-
-3. Set Up Virtual Environment:
-    ```bash
-    python -m venv venv
-    source venv/bin/activate 
-    ```
-4. Install Dependencies:
-    ```bash
-    sudo apt-get install -y ffmpeg build-essential
-    curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --no-modify-path --profile minimal
-    . "$HOME/.cargo/env"
-    export CC=/usr/bin/gcc
-    export ENV CXX=/usr/bin/g++
-    ```
-    ```bash
-    pip install --no-cache-dir --upgrade pip setuptools psutil setuptools-rust torch==2.6.0
-    pip install --no-cache-dir flash-attn --no-build-isolation
-    ```
-
-    ```bash
-    pip install -r requirements.txt
-    ```
-    
-4. Set Environment Variable:
-    Create a .env file in the root directory and add your API key:
-    plaintext
-    ```bash
-    API_KEY=your_secret_key
-    ```
-
-5. Running the Server
-    - Start the Server:
-    ```bash
-        python src/server/main.py --host 0.0.0.0 --port 7860 --config config_two
-    ```
-
-    - The server starts with models loaded on start
-    - Access the interactive API docs at http://localhost:7860/docs.
-    
-    - (Optional) Load All Models:
-    Preload all models (LLM, Translation, TTS, VLM, ASR) with:
-    - 
-    ```bash
-    curl -X POST "http://localhost:7860/load_all_models" -H "X-API-Key: your_secret_key"
-    ```
-
-- Usage
-    - Endpoints
-        - All endpoints require the X-API-Key header with the value from your .env file.
-
-    -    Health Check: GET /health
-    ```bash
-        curl "http://localhost:7860/health"
-    ```
-    - Response: 
-        ```bash 
-        {"status": "healthy", "model": "Qwen/Qwen2.5-3B-Instruct"}
-        ```
-    
-    - Text-to-Speech: POST /v1/audio/speech
-    ```  bash
-    curl -X POST "http://localhost:7860/v1/audio/speech" -H "X-API-Key: your_secret_key" -H "Content-Type: application/json" -d '{"input": "ನಮಸ್ಕಾರ", "voice": "Female voice", "model": "ai4bharat/indic-parler-tts", "response_format": "mp3"}' --output speech.mp3
-    ```
-    - Chat: POST /chat
-    ``` bash
-        curl -X POST "http://localhost:7860/chat" -H "X-API-Key: your_secret_key" -H "Content-Type: application/json" -d '{"prompt": "ನೀವು ಹೇಗಿದ್ದೀರಿ?"}'
-    ```
-
-    - Response: 
-            ```{"response": "<Kannada response>"}```
-    - Image Captioning: POST /caption/
-    ```bash
-        curl -X POST "http://localhost:7860/caption/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "length=short"
-    ```
-    - Response:``` {"caption": "<short caption>"}```
-    - Visual Query: POST /visual_query/
-    ```bash
-        curl -X POST "http://localhost:7860/visual_query/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "query=What is this?"
-    ```
-    - Response: ```{"answer": "<answer>"}```
-    - Object Detection: POST /detect/
-    ```bash
-        curl -X POST "http://localhost:7860/detect/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "object_type=face"
-    ```
-    - Response: ```{"objects": [<list of detected objects>]}```
-    - Object Pointing: POST /point/
-    ```bash
-
-        curl -X POST "http://localhost:7860/point/" -H "X-API-Key: your_secret_key" -F "file=@image.jpg" -F "object_type=person"
-    ```
-    - Response: ```{"points": [<list of points>]}```
-    - Transcription: POST /transcribe/
-    ```bash
-    curl -X POST "http://localhost:7860/transcribe/?language=kannada" -H "X-API-Key: your_secret_key" -F "file=@audio.wav"
-    ```
-    - Response: ```{"text": "<transcribed text>"}```
-    - Batch Transcription: POST /transcribe_batch/
-    ```bash
-    curl -X POST "http://localhost:7860/transcribe_batch/?language=kannada" -H "X-API-Key: your_secret_key" -F "files=@audio1.wav" -F "files=@audio2.mp3"
-    ```
-    - Response: ```{"transcriptions": ["<text1>", "<text2>"]}```
-
-- Notes
-    - Lazy Loading: Models load on first use or via /load_all_models. Expect a delay on the first request for each model type.
-    Supported Languages: ASR supports multiple Indian languages (e.g., kannada, hindi, tamil); see models/asr.py for the full list.
-    Logs: Check dhwani_api.log for detailed logs (rotated at 10MB, 5 backups).
-    Performance: Use a GPU with flash-attn installed for faster TTS and ASR inference.
-
-- Troubleshooting
-
-    - Module Errors: Ensure all dependencies are installed. Re-run pip install if needed.
-    FFmpeg Not Found: Install FFmpeg and ensure it’s in your PATH.
-    Permission Denied: Run with sudo if accessing restricted ports (e.g., < 1024).
-

dwani-0.1.1/src/dwani.egg-info/SOURCES.txt

@@ -1,16 +0,0 @@
-LICENSE
-README.md
-pyproject.toml
-src/__init__.py
-src/asr.py
-src/audio.py
-src/chat.py
-src/client.py
-src/docs.py
-src/exceptions.py
-src/vision.py
-src/dwani.egg-info/PKG-INFO
-src/dwani.egg-info/SOURCES.txt
-src/dwani.egg-info/dependency_links.txt
-src/dwani.egg-info/requires.txt
-src/dwani.egg-info/top_level.txt

dwani-0.1.1/src/dwani.egg-info/requires.txt

@@ -1,6 +0,0 @@
-fastapi>=0.95.0
-uvicorn[standard]>=0.22.0
-pydantic>=2.0.0
-requests>=2.25.0
-python-multipart>=0.0.5
-pydantic-settings>=2.0.0

dwani-0.1.1/src/dwani.egg-info/top_level.txt

@@ -1,8 +0,0 @@
-__init__
-asr
-audio
-chat
-client
-docs
-exceptions
-vision