dwani 0.1.6__py3-none-any.whl → 0.1.7__py3-none-any.whl

dwani/asr.py

@@ -1,10 +1,44 @@
 from .exceptions import DhwaniAPIError
 import requests
+
+# Allowed languages (case-sensitive for display, but we'll handle case-insensitively)
+ALLOWED_LANGUAGES = [
+    "Assamese",
+    "Bengali",
+    "Gujarati",
+    "Hindi",
+    "Kannada",
+    "Malayalam",
+    "Marathi",
+    "Odia",
+    "Punjabi",
+    "Tamil",
+    "Telugu"
+]
+
+def validate_language(language):
+    """Validate that the provided language is in the allowed list (case-insensitive)."""
+    # Create a case-insensitive mapping of allowed languages
+    language_map = {lang.lower(): lang for lang in ALLOWED_LANGUAGES}
+    # Check if the lowercase version of the input language is in the map
+    if language.lower() not in language_map:
+        raise ValueError(
+            f"Unsupported language: {language}. Supported languages: {ALLOWED_LANGUAGES}"
+        )
+    # Return the original case from ALLOWED_LANGUAGES for consistency
+    return language_map[language.lower()]
+
 def asr_transcribe(client, file_path, language):
+    # Validate the language input (case-insensitive)
+    validated_language = validate_language(language)
+    
+    # Convert language to lowercase for the API request
+    api_language = validated_language.lower()
+    
     with open(file_path, "rb") as f:
         files = {"file": f}
         resp = requests.post(
-            f"{client.api_base}/v1/transcribe/?language={language}",
+            f"{client.api_base}/v1/transcribe/?language={api_language}",
             headers=client._headers(),
             files=files
         )
@@ -16,5 +50,4 @@ class ASR:
     @staticmethod
     def transcribe(*args, **kwargs):
         from . import _get_client
-        return _get_client().transcribe(*args, **kwargs)
-
+        return _get_client().transcribe(*args, **kwargs)

dwani/chat.py

@@ -1,12 +1,50 @@
 from .exceptions import DhwaniAPIError
 import requests
 
+# Language options mapping
+language_options = [
+    ("English", "eng_Latn"),
+    ("Kannada", "kan_Knda"),
+    ("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") 
+]
+
+# Create a dictionary for language name to code mapping
+lang_name_to_code = {name.lower(): code for name, code in language_options}
+lang_code_to_code = {code: code for _, code in language_options}
+
+def normalize_language(lang):
+    """Convert language input (name or code) to language code."""
+    lang = lang.strip()
+    # Check if input is a language name (case-insensitive)
+    lang_lower = lang.lower()
+    if lang_lower in lang_name_to_code:
+        return lang_name_to_code[lang_lower]
+    # Check if input is a language code
+    if lang in lang_code_to_code:
+        return lang_code_to_code[lang]
+    # Raise error if language is not supported
+    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 chat_create(client, prompt, src_lang, tgt_lang, **kwargs):
+    # Normalize source and target languages
+    src_lang_code = normalize_language(src_lang)
+    tgt_lang_code = normalize_language(tgt_lang)
+    
     url = f"{client.api_base}/v1/indic_chat"
     payload = {
         "prompt": prompt,
-        "src_lang": src_lang,
-        "tgt_lang": tgt_lang
+        "src_lang": src_lang_code,
+        "tgt_lang": tgt_lang_code
     }
     payload.update(kwargs)
     resp = requests.post(
@@ -22,4 +60,4 @@ 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)
+        return _get_client().chat(prompt, src_lang, tgt_lang, **kwargs)

dwani/docs.py

@@ -1,13 +1,49 @@
 import requests
 from .exceptions import DhwaniAPIError
 
+# Language options mapping
+language_options = [
+    ("English", "eng_Latn"),
+    ("Kannada", "kan_Knda"),
+    ("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") 
+]
+
+# Create dictionaries for language name to code and code to code mapping
+lang_name_to_code = {name.lower(): code for name, code in language_options}
+lang_code_to_code = {code: code for _, code in language_options}
+
+def normalize_language(lang):
+    """Convert language input (name or code) to language code."""
+    lang = lang.strip()
+    # Check if input is a language name (case-insensitive)
+    lang_lower = lang.lower()
+    if lang_lower in lang_name_to_code:
+        return lang_name_to_code[lang_lower]
+    # Check if input is a language code
+    if lang in lang_code_to_code:
+        return lang_code_to_code[lang]
+    # Raise error if language is not supported
+    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 document_ocr(client, file_path, language=None):
     """OCR a document (image/PDF) and return extracted text."""
+    data = {}
+    if language:
+        # Normalize the language input
+        data["language"] = normalize_language(language)
+    
     with open(file_path, "rb") as f:
         files = {"file": f}
-        data = {}
-        if language:
-            data["language"] = language
         resp = requests.post(
             f"{client.api_base}/v1/document/ocr",
             headers=client._headers(),
@@ -20,14 +56,18 @@ def document_ocr(client, file_path, language=None):
 
 def document_summarize(client, file_path, page_number=1, src_lang="eng_Latn", tgt_lang="kan_Knda"):
     """Summarize a PDF document with language and page number options."""
+    # Normalize source and target languages
+    src_lang_code = normalize_language(src_lang)
+    tgt_lang_code = normalize_language(tgt_lang)
+    
     url = f"{client.api_base}/v1/indic-summarize-pdf"
     headers = client._headers()
     with open(file_path, "rb") as f:
         files = {"file": (file_path, f, "application/pdf")}
         data = {
             "page_number": str(page_number),
-            "src_lang": src_lang,
-            "tgt_lang": tgt_lang
+            "src_lang": src_lang_code,
+            "tgt_lang": tgt_lang_code
         }
         resp = requests.post(
             url,
@@ -39,18 +79,20 @@ def document_summarize(client, file_path, page_number=1, src_lang="eng_Latn", tg
         raise DhwaniAPIError(resp)
     return resp.json()
 
-
 def extract(client, file_path, page_number, src_lang, tgt_lang):
     """
     Extract and translate text from a document (image/PDF) using query parameters.
     """
+    # Normalize source and target languages
+    src_lang_code = normalize_language(src_lang)
+    tgt_lang_code = normalize_language(tgt_lang)
+    
     # Build the URL with query parameters
     url = (
         f"{client.api_base}/v1/indic-extract-text/"
-        f"?page_number={page_number}&src_lang={src_lang}&tgt_lang={tgt_lang}"
+        f"?page_number={page_number}&src_lang={src_lang_code}&tgt_lang={tgt_lang_code}"
     )
     headers = client._headers()
-    # 'requests' handles multipart/form-data automatically
     with open(file_path, "rb") as f:
         files = {"file": (file_path, f, "application/pdf")}
         resp = requests.post(
@@ -71,6 +113,10 @@ def doc_query(
     tgt_lang="kan_Knda"
 ):
     """Query a document with a custom prompt and language options."""
+    # Normalize source and target languages
+    src_lang_code = normalize_language(src_lang)
+    tgt_lang_code = normalize_language(tgt_lang)
+    
     url = f"{client.api_base}/v1/indic-custom-prompt-pdf"
     headers = client._headers()
     with open(file_path, "rb") as f:
@@ -78,8 +124,8 @@ def doc_query(
         data = {
             "page_number": str(page_number),
             "prompt": prompt,
-            "source_language": src_lang,
-            "target_language": tgt_lang
+            "source_language": src_lang_code,
+            "target_language": tgt_lang_code
         }
         resp = requests.post(
             url,
@@ -91,7 +137,6 @@ def doc_query(
         raise DhwaniAPIError(resp)
     return resp.json()
 
-
 def doc_query_kannada(
     client, 
     file_path, 
@@ -101,18 +146,20 @@ def doc_query_kannada(
     language=None
 ):
     """Summarize a document (image/PDF/text) with custom prompt and language."""
+    # Normalize source language and optional language parameter
+    src_lang_code = normalize_language(src_lang)
+    data = {
+        "page_number": str(page_number),
+        "prompt": prompt,
+        "src_lang": src_lang_code,
+    }
+    if language:
+        data["language"] = normalize_language(language)
+    
     url = f"{client.api_base}/v1/indic-custom-prompt-kannada-pdf"
     headers = client._headers()
-    # 'requests' will handle multipart/form-data automatically
     with open(file_path, "rb") as f:
         files = {"file": (file_path, f, "application/pdf")}
-        data = {
-            "page_number": str(page_number),
-            "prompt": prompt,
-            "src_lang": src_lang,
-        }
-        if language:
-            data["language"] = language
         resp = requests.post(
             url,
             headers=headers,
@@ -123,8 +170,6 @@ def doc_query_kannada(
         raise DhwaniAPIError(resp)
     return resp.json()
 
-
-
 class Documents:
     @staticmethod
     def ocr(file_path, language=None):
@@ -135,14 +180,17 @@ class Documents:
     def summarize(*args, **kwargs):
         from . import _get_client
         return _get_client().document_summarize(*args, **kwargs)
+    
     @staticmethod
     def run_extract(*args, **kwargs):
         from . import _get_client
         return _get_client().extract(*args, **kwargs)
+    
     @staticmethod
     def run_doc_query(*args, **kwargs):
         from . import _get_client
         return _get_client().doc_query(*args, **kwargs)
+    
     @staticmethod
     def run_doc_query_kannada(*args, **kwargs):
         from . import _get_client

dwani/translate.py

@@ -1,12 +1,50 @@
 from .exceptions import DhwaniAPIError
 import requests
 
+# Language options mapping
+language_options = [
+    ("English", "eng_Latn"),
+    ("Kannada", "kan_Knda"),
+    ("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") 
+]
+
+# Create dictionaries for language name to code and code to code mapping
+lang_name_to_code = {name.lower(): code for name, code in language_options}
+lang_code_to_code = {code: code for _, code in language_options}
+
+def normalize_language(lang):
+    """Convert language input (name or code) to language code."""
+    lang = lang.strip()
+    # Check if input is a language name (case-insensitive)
+    lang_lower = lang.lower()
+    if lang_lower in lang_name_to_code:
+        return lang_name_to_code[lang_lower]
+    # Check if input is a language code
+    if lang in lang_code_to_code:
+        return lang_code_to_code[lang]
+    # Raise error if language is not supported
+    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 run_translate(client, sentences, src_lang, tgt_lang, **kwargs):
+    # Normalize source and target languages
+    src_lang_code = normalize_language(src_lang)
+    tgt_lang_code = normalize_language(tgt_lang)
+    
     url = f"{client.api_base}/v1/translate"
     payload = {
         "sentences": sentences,
-        "src_lang": src_lang,
-        "tgt_lang": tgt_lang
+        "src_lang": src_lang_code,
+        "tgt_lang": tgt_lang_code
     }
     payload.update(kwargs)
     resp = requests.post(
@@ -22,5 +60,4 @@ class Translate:
     @staticmethod
     def run_translate(sentences, src_lang, tgt_lang, **kwargs):
         from . import _get_client
-        return _get_client().translate(sentences, src_lang, tgt_lang, **kwargs)
-        
+        return _get_client().translate(sentences, src_lang, tgt_lang, **kwargs)

dwani/vision.py

@@ -1,15 +1,53 @@
 from .exceptions import DhwaniAPIError
 import requests
+
+# Language options mapping
+language_options = [
+    ("English", "eng_Latn"),
+    ("Kannada", "kan_Knda"),
+    ("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") 
+]
+
+# Create dictionaries for language name to code and code to code mapping
+lang_name_to_code = {name.lower(): code for name, code in language_options}
+lang_code_to_code = {code: code for _, code in language_options}
+
+def normalize_language(lang):
+    """Convert language input (name or code) to language code."""
+    lang = lang.strip()
+    # Check if input is a language name (case-insensitive)
+    lang_lower = lang.lower()
+    if lang_lower in lang_name_to_code:
+        return lang_name_to_code[lang_lower]
+    # Check if input is a language code
+    if lang in lang_code_to_code:
+        return lang_code_to_code[lang]
+    # Raise error if language is not supported
+    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 vision_caption(client, file_path, query="describe the image", src_lang="eng_Latn", tgt_lang="kan_Knda"):
+    # Normalize source and target languages
+    src_lang_code = normalize_language(src_lang)
+    tgt_lang_code = normalize_language(tgt_lang)
+    
     # Build the endpoint using the client's api_base
     url = (
         f"{client.api_base}/v1/indic_visual_query"
-        f"?src_lang={src_lang}&tgt_lang={tgt_lang}"
+        f"?src_lang={src_lang_code}&tgt_lang={tgt_lang_code}"
     )
     headers = {
         **client._headers(),
         "accept": "application/json"
-        # Note: 'Content-Type' will be set automatically by requests when using 'files'
     }
     with open(file_path, "rb") as f:
         files = {"file": (file_path, f, "image/png")}
@@ -28,4 +66,4 @@ class Vision:
     @staticmethod
     def caption(*args, **kwargs):
         from . import _get_client
-        return _get_client().caption(*args, **kwargs)
+        return _get_client().caption(*args, **kwargs)

{dwani-0.1.6.dist-info → dwani-0.1.7.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dwani
-Version: 0.1.6
+Version: 0.1.7
 Summary: Multimodal API for Indian languages (Chat, Vision, TTS, ASR, Translate, Docs)
 Author-email: sachin <python@dwani.ai>
 License: MIT License
@@ -42,6 +42,9 @@ Dynamic: license-file
 pip install dwani
 ```
 
+### Languages supported
+    - Assamese, Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi, Odia, Punjabi, Tamil, Telugu
+
 ### Setup the credentials
 ```python
 import dwani

dwani-0.1.7.dist-info/RECORD

@@ -0,0 +1,14 @@
+dwani/__init__.py,sha256=ldO5OND7DvJlbxaQ0R57Cc73jJTnCSslDDt4I4r-Op8,1895
+dwani/asr.py,sha256=3LYrLOaMhc5eXKFSoi63C8KAvwZI2NcuO25pwTfSVe0,1692
+dwani/audio.py,sha256=Q9vw4uBxGy1vQzmiZjZGrY8hkAEQNkGhjz5OcnpFEQQ,888
+dwani/chat.py,sha256=dQCl8lLQczwnAsvYlTZowd471ktRVZcW3w8gZ5Wpzms,2097
+dwani/client.py,sha256=OrnwqxBQMfEZ1iQEleFigNujiZve3ox53yv5aSmB3iQ,2849
+dwani/docs.py,sha256=EO41opJwfDFsNmH6nQl-HOsyWravCnj1f5ZDgxSZECI,6323
+dwani/exceptions.py,sha256=qEN5ukqlnN7v-kHNEnISWFMpPMt6uTft9mPsTXJ4LVA,227
+dwani/translate.py,sha256=IJiKrYIfwdJKc_PjlZKVRAwzpQDst_2MF_B_huxid_E,2185
+dwani/vision.py,sha256=wN7WkMRVmLrZnBJxnam7vihTXWGlWJ4JqXgyrp-tbrg,2330
+dwani-0.1.7.dist-info/licenses/LICENSE,sha256=IAD8tbwWZbPWHXgYjabHoMv0aaUzZUYzYiEbfhTCisY,1070
+dwani-0.1.7.dist-info/METADATA,sha256=ggOY4wss1nwN6PzhB0BZG5332emrP6C76qwjaK2mzSs,4810
+dwani-0.1.7.dist-info/WHEEL,sha256=Nw36Djuh_5VDukK0H78QzOX-_FQEo6V37m3nkm96gtU,91
+dwani-0.1.7.dist-info/top_level.txt,sha256=AM5EhkyuO_EXQFR9JIxEV6tAYMCCyc-a1dLifpCGBUk,6
+dwani-0.1.7.dist-info/RECORD,,

{dwani-0.1.6.dist-info → dwani-0.1.7.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.4.0)
+Generator: setuptools (80.7.1)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

dwani-0.1.6.dist-info/RECORD

@@ -1,14 +0,0 @@
-dwani/__init__.py,sha256=ldO5OND7DvJlbxaQ0R57Cc73jJTnCSslDDt4I4r-Op8,1895
-dwani/asr.py,sha256=Y5Mbv1KsvhfXNZacMZycUHPn79NRQC-XPH0j9SYPUSY,590
-dwani/audio.py,sha256=Q9vw4uBxGy1vQzmiZjZGrY8hkAEQNkGhjz5OcnpFEQQ,888
-dwani/chat.py,sha256=WFuEShNd4nd6KUbIZTkm3eyPGoP33GepOLJax86nNn8,720
-dwani/client.py,sha256=OrnwqxBQMfEZ1iQEleFigNujiZve3ox53yv5aSmB3iQ,2849
-dwani/docs.py,sha256=s7lgw9tfA0IlItAMlwnPuuxMEmS1ls9ygEIgjuUKlcg,4550
-dwani/exceptions.py,sha256=qEN5ukqlnN7v-kHNEnISWFMpPMt6uTft9mPsTXJ4LVA,227
-dwani/translate.py,sha256=SH8daC0MLudM3GX3ezoWufNmh1u1Ym4LELiUlWHeyeM,791
-dwani/vision.py,sha256=JtMSS0hI-8gxtLugVP10__enZsPPy8jheMyWXvrGrdw,1015
-dwani-0.1.6.dist-info/licenses/LICENSE,sha256=IAD8tbwWZbPWHXgYjabHoMv0aaUzZUYzYiEbfhTCisY,1070
-dwani-0.1.6.dist-info/METADATA,sha256=A3XIToek1Gj1y4rX8Ipd7-6fEa2kShIjUVl6ZD38j4o,4685
-dwani-0.1.6.dist-info/WHEEL,sha256=DnLRTWE75wApRYVsjgc6wsVswC54sMSJhAEd4xhDpBk,91
-dwani-0.1.6.dist-info/top_level.txt,sha256=AM5EhkyuO_EXQFR9JIxEV6tAYMCCyc-a1dLifpCGBUk,6
-dwani-0.1.6.dist-info/RECORD,,