dwani 0.1.21__tar.gz → 0.1.23__tar.gz

{dwani-0.1.21 → dwani-0.1.23}/PKG-INFO

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: dwani
-Version: 0.1.21
-Summary: Multimodal API for Indian + European languages (Chat, Vision, TTS, ASR, Translate, Docs)
+Version: 0.1.23
+Summary: Document Analytics + Multimodal API for Indian + European languages (Chat, Vision, TTS, ASR, Translate, Docs)
 Author-email: sachin <python@dwani.ai>
 License: MIT License
         
@@ -25,9 +25,9 @@ License: MIT License
         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
+Project-URL: Homepage, https://github.com/dwani-ai/dwani-python-sdk
+Project-URL: Source, https://github.com/dwani-ai/dwani-python-sdk
+Project-URL: Issues, https://github.com/dwani-ai/dwani-python-sdk/issues
 Requires-Python: >=3.8
 Description-Content-Type: text/markdown
 License-File: LICENSE
@@ -48,6 +48,12 @@ Dynamic: license-file
 pip install --upgrade dwani
 ```
 
+### Model Supported
+- Text
+  - gpt-oss , gemma3
+- Vision
+  - gemma3
+
 ### Languages supported
 - Indian
   - Assamese, Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi Odia, Punjabi, Tamil, Telugu

{dwani-0.1.21 → dwani-0.1.23}/README.md

@@ -11,6 +11,12 @@
 pip install --upgrade dwani
 ```
 
+### Model Supported
+- Text
+  - gpt-oss , gemma3
+- Vision
+  - gemma3
+
 ### Languages supported
 - Indian
   - Assamese, Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi Odia, Punjabi, Tamil, Telugu

{dwani-0.1.21 → dwani-0.1.23}/dwani/__init__.py

@@ -40,7 +40,9 @@ class vision:
     @staticmethod
     def caption_direct(file_path, query="describe the image", model="gemma3", system_prompt=""):
         return _get_client().caption_direct(file_path, query, model, system_prompt)
-
+    @staticmethod
+    def ocr_image(file_path, model="gemma3"):
+        return _get_client().ocr_image(file_path, model)
 class asr:
     @staticmethod
     def transcribe(file_path, language="kannada"):

{dwani-0.1.21 → dwani-0.1.23}/dwani/chat.py

@@ -37,6 +37,10 @@ def normalize_language(lang):
     raise ValueError(f"Unsupported language: {lang}. Supported languages: {supported_langs}")
 
 def chat_direct(client, prompt, model="gemma3", system_prompt=""):
+    valid_models = ["gemma3", "qwen3", "gpt-oss", "sarvam-m"]
+    if model not in valid_models:
+        raise ValueError(f"Unsupported model: {model}. Supported models: {valid_models}")
+
     url = f"{client.api_base}/v1/chat_direct"
     payload = {
         "prompt": prompt,
@@ -55,7 +59,7 @@ def chat_direct(client, prompt, model="gemma3", system_prompt=""):
 
 def chat_create(client, prompt, src_lang, tgt_lang, model="gemma3"):
     # Validate model
-    valid_models = ["gemma3", "qwen3", "deepseek-r1", "sarvam-m"]
+    valid_models = ["gemma3", "qwen3", "gpt-oss", "sarvam-m"]
     if model not in valid_models:
         raise ValueError(f"Unsupported model: {model}. Supported models: {valid_models}")
     

{dwani-0.1.21 → dwani-0.1.23}/dwani/client.py

@@ -35,6 +35,10 @@ class DwaniClient:
         from .vision import vision_caption
         return vision_caption(self, file_path=file_path, query=query, src_lang=src_lang, tgt_lang=tgt_lang, model=model)
 
+    def ocr_image(self, file_path, model="gemma3"):
+        from .vision import ocr_image
+        return ocr_image(self, file_path=file_path, model=model)
+
     def caption_direct(self, file_path, query="describe the image", model="gemma3", system_prompt=""):
         from .vision import vision_direct
         return vision_direct(self, file_path=file_path, query=query, model=model, system_prompt=system_prompt)

{dwani-0.1.21 → dwani-0.1.23}/dwani/docs.py

@@ -10,7 +10,14 @@ logger = logging.getLogger(__name__)
 language_options = [
     ("English", "eng_Latn"),
     ("Kannada", "kan_Knda"),
-    ("Hindi", "hin_Deva"),
+    ("Hindi", "hin_Deva"), 
+    ("Assamese", "asm_Beng"),
+    ("Bengali", "ben_Beng"),
+    ("Gujarati", "guj_Gujr"),
+    ("Malayalam", "mal_Mlym"),
+    ("Marathi", "mar_Deva"),
+    ("Odia", "ory_Orya"),
+    ("Punjabi", "pan_Guru"),
     ("Tamil", "tam_Taml"),
     ("Telugu", "tel_Telu"),
     ("German", "deu_Latn") 
@@ -53,6 +60,8 @@ def document_ocr_all(client, file_path, model="gemma3"):
         try:
             resp = requests.post(
                 f"{client.api_base}/v1/extract-text-all",
+#TODO - test -chunk
+#                f"{client.api_base}/v1/extract-text-all-chunk",
                 headers=client._headers(),
                 files=files,
                 data=data,

{dwani-0.1.21 → dwani-0.1.23}/dwani/vision.py

@@ -36,6 +36,28 @@ def normalize_language(lang):
     supported_langs = list(lang_name_to_code.keys()) + list(lang_code_to_code.keys())
     raise ValueError(f"Unsupported language: {lang}. Supported languages: {supported_langs}")
 
+def ocr_image(client, file_path, model="gemma3"):
+    url = (
+        f"{client.api_base}/v1/ocr"
+        f"?model={model}"
+    )
+    headers = {
+        **client._headers(),
+        "accept": "application/json"
+    }
+    with open(file_path, "rb") as f:
+        files = {"file": (file_path, f, "image/png")}
+        resp = requests.post(
+            url,
+            headers=headers,
+            files=files,
+            timeout=90
+        )
+    if resp.status_code != 200:
+        raise DwaniAPIError(resp)
+    return resp.json()
+
+
 def vision_direct(client, file_path, query="describe this image", model="gemma3", system_prompt=""):
     url = (
         f"{client.api_base}/v1/visual_query_direct"
@@ -101,3 +123,7 @@ class Vision:
     def caption_direct(file_path, query="describe the image", model="gemma3", system_prompt=""):
         from . import _get_client
         return _get_client().caption_direct(file_path, query, model, system_prompt)
+    @staticmethod
+    def ocr_image(file_path, model="gemma3"):
+        from . import _get_client
+        return _get_client().ocr_image(file_path, model)

{dwani-0.1.21 → dwani-0.1.23}/pyproject.toml

@@ -5,8 +5,8 @@ build-backend = "setuptools.build_meta"
 [project]
 name = "dwani"
 
-version = "0.1.21"
-description = "Multimodal API for Indian + European languages (Chat, Vision, TTS, ASR, Translate, Docs)"
+version = "0.1.23"
+description = "Document Analytics + Multimodal API for Indian + European languages (Chat, Vision, TTS, ASR, Translate, Docs)"
 authors = [
     { name="sachin", email="python@dwani.ai" }
 ]
@@ -20,6 +20,6 @@ dependencies = [
 ]
 
 [project.urls]
-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"
+Homepage = "https://github.com/dwani-ai/dwani-python-sdk"
+Source = "https://github.com/dwani-ai/dwani-python-sdk"
+Issues = "https://github.com/dwani-ai/dwani-python-sdk/issues"