lumera 0.10.1__tar.gz → 0.10.2__tar.gz

{lumera-0.10.1 → lumera-0.10.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lumera
-Version: 0.10.1
+Version: 0.10.2
 Summary: SDK for building on Lumera platform
 Requires-Python: >=3.11
 Requires-Dist: requests

{lumera-0.10.1 → lumera-0.10.2}/lumera/__init__.py

@@ -13,7 +13,7 @@ except PackageNotFoundError:
     __version__ = "unknown"  # Not installed (e.g., running from source)
 
 # Import new modules (as modules, not individual functions)
-from . import automations, email, exceptions, integrations, llm, locks, pb, storage, webhooks
+from . import automations, documents, email, exceptions, integrations, llm, locks, pb, storage, webhooks
 from ._utils import (
     LumeraAPIError,
     RecordNotUniqueError,
@@ -102,6 +102,7 @@ __all__ = [
     "LockHeldError",
     # New modules (use as lumera.pb, lumera.storage, etc.)
     "automations",
+    "documents",
     "email",
     "pb",
     "storage",

lumera-0.10.2/lumera/documents.py

@@ -0,0 +1,127 @@
+"""
+Document text extraction via OpenAI Responses API.
+
+Supports images and PDFs.
+
+Example:
+    >>> from lumera import documents
+    >>> text = documents.extract_text("invoice.pdf")
+
+    # Or from bytes (no file needed):
+    >>> text = documents.extract_text_from_bytes(pdf_bytes, "application/pdf")
+"""
+
+from __future__ import annotations
+
+import base64
+import mimetypes
+from pathlib import Path
+
+__all__ = ["extract_text", "extract_text_from_bytes"]
+
+_DEFAULT_MODEL = "gpt-5-mini"
+_DEFAULT_PROMPT = "Extract all text from this document. Return only the extracted text."
+
+
+def _get_mime_type(file_path: str) -> str:
+    """Get MIME type for a file."""
+    mime, _ = mimetypes.guess_type(file_path)
+    if mime:
+        return mime
+    ext = Path(file_path).suffix.lower()
+    return {
+        ".png": "image/png",
+        ".jpg": "image/jpeg",
+        ".jpeg": "image/jpeg",
+        ".gif": "image/gif",
+        ".webp": "image/webp",
+        ".pdf": "application/pdf",
+    }.get(ext, "application/octet-stream")
+
+
+def extract_text_from_bytes(
+    data: bytes,
+    mime_type: str,
+    *,
+    filename: str = "document",
+    prompt: str = _DEFAULT_PROMPT,
+    model: str = _DEFAULT_MODEL,
+) -> str:
+    """Extract text from document bytes using OpenAI.
+
+    Args:
+        data: Raw file bytes
+        mime_type: MIME type (e.g., "application/pdf", "image/png")
+        filename: Optional filename for context
+        prompt: What to ask the LLM
+        model: Model to use (default: gpt-5-mini)
+
+    Returns:
+        Extracted text
+
+    Example:
+        >>> import requests
+        >>> resp = requests.get(presigned_url)
+        >>> text = documents.extract_text_from_bytes(
+        ...     resp.content,
+        ...     "application/pdf",
+        ...     filename="invoice.pdf"
+        ... )
+    """
+    from . import llm
+
+    b64 = base64.b64encode(data).decode("utf-8")
+    data_url = f"data:{mime_type};base64,{b64}"
+    client = llm.get_provider().client
+
+    response = client.responses.create(
+        model=model,
+        input=[
+            {
+                "role": "user",
+                "content": [
+                    {
+                        "type": "input_file",
+                        "filename": filename,
+                        "file_data": data_url,
+                    },
+                    {
+                        "type": "input_text",
+                        "text": prompt,
+                    },
+                ],
+            },
+        ],
+    )
+    return response.output_text or ""
+
+
+def extract_text(
+    file_path: str,
+    *,
+    prompt: str = _DEFAULT_PROMPT,
+    model: str = _DEFAULT_MODEL,
+) -> str:
+    """Extract text from a document file using OpenAI.
+
+    Args:
+        file_path: Path to image or PDF
+        prompt: What to ask the LLM
+        model: Model to use (default: gpt-5-mini)
+
+    Returns:
+        Extracted text
+    """
+    with open(file_path, "rb") as f:
+        data = f.read()
+
+    mime = _get_mime_type(file_path)
+    filename = Path(file_path).name
+
+    return extract_text_from_bytes(
+        data,
+        mime,
+        filename=filename,
+        prompt=prompt,
+        model=model,
+    )

{lumera-0.10.1 → lumera-0.10.2}/lumera/llm.py

@@ -16,7 +16,7 @@ Configuration:
 
 Example:
     >>> from lumera import llm
-    >>> response = llm.complete("What is 2+2?", model="gpt-5.2-mini")
+    >>> response = llm.complete("What is 2+2?", model="gpt-5-mini")
     >>> print(response["content"])
 """
 
@@ -135,14 +135,14 @@ class OpenAIProvider(LLMProvider):
     # Model aliases for convenience
     MODEL_ALIASES: dict[str, str] = {
         "gpt-5.2": "gpt-5.2",
-        "gpt-5.2-mini": "gpt-5.2-mini",
-        "gpt-5.2-nano": "gpt-5.2-nano",
+        "gpt-5-mini": "gpt-5-mini",
+        "gpt-5-nano": "gpt-5-nano",
         # Embedding models
         "text-embedding-3-small": "text-embedding-3-small",
         "text-embedding-3-large": "text-embedding-3-large",
     }
 
-    DEFAULT_CHAT_MODEL = "gpt-5.2-mini"
+    DEFAULT_CHAT_MODEL = "gpt-5-mini"
     DEFAULT_EMBEDDING_MODEL = "text-embedding-3-small"
     DEFAULT_PROVIDER_NAME = "openai"
 
@@ -374,7 +374,7 @@ def set_provider(provider: LLMProvider | str, **kwargs: Unpack[ProviderConfig])
 def complete(
     prompt: str,
     *,
-    model: str = "gpt-5.2-mini",
+    model: str = "gpt-5-mini",
     temperature: float = 0.7,
     max_tokens: int | None = None,
     system_prompt: str | None = None,
@@ -384,7 +384,7 @@ def complete(
 
     Args:
         prompt: User prompt/question
-        model: Model to use (default: gpt-5.2-mini)
+        model: Model to use (default: gpt-5-mini)
         temperature: Sampling temperature 0.0 to 2.0 (default: 0.7)
         max_tokens: Max tokens in response (None = model default)
         system_prompt: Optional system message to set behavior
@@ -397,7 +397,7 @@ def complete(
         >>> response = llm.complete(
         ...     prompt="Classify this deposit: ...",
         ...     system_prompt="You are an expert accountant.",
-        ...     model="gpt-5.2-mini",
+        ...     model="gpt-5-mini",
         ...     json_mode=True
         ... )
         >>> data = json.loads(response["content"])
@@ -416,7 +416,7 @@ def complete(
 def chat(
     messages: list[Message],
     *,
-    model: str = "gpt-5.2-mini",
+    model: str = "gpt-5-mini",
     temperature: float = 0.7,
     max_tokens: int | None = None,
     json_mode: bool = False,
@@ -425,7 +425,7 @@ def chat(
 
     Args:
         messages: Conversation history with role and content
-        model: Model to use (default: gpt-5.2-mini)
+        model: Model to use (default: gpt-5-mini)
         temperature: Sampling temperature 0.0 to 2.0 (default: 0.7)
         max_tokens: Max tokens in response (None = model default)
         json_mode: Force JSON output (default: False)

{lumera-0.10.1 → lumera-0.10.2}/lumera.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: lumera
-Version: 0.10.1
+Version: 0.10.2
 Summary: SDK for building on Lumera platform
 Requires-Python: >=3.11
 Requires-Dist: requests

{lumera-0.10.1 → lumera-0.10.2}/lumera.egg-info/SOURCES.txt

@@ -2,6 +2,7 @@ pyproject.toml
 lumera/__init__.py
 lumera/_utils.py
 lumera/automations.py
+lumera/documents.py
 lumera/email.py
 lumera/exceptions.py
 lumera/files.py

{lumera-0.10.1 → lumera-0.10.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "lumera"
-version = "0.10.1"
+version = "0.10.2"
 description = "SDK for building on Lumera platform"
 requires-python = ">=3.11"
 dependencies = [