sunholo 0.108.0__py3-none-any.whl → 0.109.3__py3-none-any.whl

sunholo/genai/__init__.py

@@ -1,3 +1,4 @@
 from .process_funcs_cls import GenAIFunctionProcessor
 from .safety import genai_safety
-from .init import init_genai
+from .init import init_genai
+from .file_handling import download_gcs_upload_genai, construct_file_content

sunholo/genai/file_handling.py

@@ -0,0 +1,221 @@
+from ..custom_logging import log
+from ..gcs import get_bytes_from_gcs
+
+from functools import partial
+import mimetypes
+import asyncio
+import tempfile
+import re
+import os
+import traceback
+try:
+    import google.generativeai as genai
+except ImportError:
+    genai = None
+
+DOCUMENT_MIMES = [
+    'application/pdf',
+    'application/x-javascript',
+    'text/javascript',
+    'application/x-python',
+    'text/x-python',
+    'text/plain',
+    'text/html',
+    'text/css',
+    'text/md',
+    'text/csv',
+    'text/xml',
+    'text/rtf'
+    ]
+
+IMAGE_MIMES = [
+    'image/png',
+    'image/jpeg',
+    'image/webp',
+    'image/heic',
+    'image/heif',
+]
+
+VIDEO_MIMES = [
+    'video/mp4',
+    'video/mpeg',
+    'video/mov',
+    'video/avi',
+    'video/x-flv',
+    'video/mpg',
+    'video/webm',
+    'video/wmv',
+    'video/3gpp'
+]
+
+AUDIO_MIMES = [
+    'audio/wav',
+    'audio/mp3',
+    'audio/aiff',
+    'audio/aac',
+    'audio/ogg',
+    'audio/flac',    
+]
+
+ALLOWED_MIME_TYPES = set(AUDIO_MIMES + VIDEO_MIMES + IMAGE_MIMES + DOCUMENT_MIMES)
+
+# 'documents': 
+# [
+#   {'storagePath': 'users/UQcKi4u7s...dsd.png', 
+#    'url': 'https://firebasestorage.googleapis.com/v0/b/multi...', 
+#    'contentType': 'image/png', 
+#    'type': 'image', 
+#    'name': 'multivac-data-architecture.png'}, 
+#   {'storagePath': 'users/UQc...3dc59e1.jpg', 
+#   'type': 'image', 
+#   'name': 'holosun-circle.jpg', 
+#   'url': 'https://firebasestorage.googleapis.com/v0/b/multiv...', 
+#   'contentType': 'image/jpeg'}
+# ]
+
+def sanitize_file(filename):
+    # Split the filename into name and extension
+    name, extension = os.path.splitext(filename)
+    
+    # Sanitize the name by removing invalid characters and converting to lowercase
+    sanitized_name = re.sub(r'[^a-z0-9-]', '', name.lower())
+    sanitized_name = re.sub(r'^-+|-+$', '', sanitized_name)  # Remove leading or trailing dashes
+    
+    # Reattach the original extension
+    return f"{sanitized_name}"
+
+async def construct_file_content(gs_list, bucket:str):
+    """
+    Args:
+    - gs_list: a list of dicts representing files in a bucket
+       - contentType: The content type of the file on GCS
+       - storagePath: The path in the bucket
+       - name: The name of the file
+    - bucket: The bucket the files are in
+
+    """
+    
+    file_list = []
+    for element in gs_list:
+
+        the_mime_type = element.get('contentType')
+        if the_mime_type is None:
+            continue
+        if element.get('storagePath') is None:
+            continue
+        if the_mime_type in ALLOWED_MIME_TYPES:
+            file_list.append(element)
+    
+    if not file_list:
+        return {"role": "user", "parts": [{"text": "No eligible contentTypes were found"}]}
+
+    content = []
+    
+    # Loop through the valid files and process them
+    tasks = []
+    for file_info in file_list:
+        img_url = f"gs://{bucket}/{file_info['storagePath']}"
+        mime_type = file_info['contentType']
+        name = sanitize_file(file_info['name'])
+        log.info(f"Processing {name=}")
+        try:
+            myfile = genai.get_file(name)
+            content.append(
+                {"role": "user", "parts": [
+                    {"file_data": myfile}, 
+                    {"text": f"You have been given the ability to work with file '{name}' with {mime_type=}."}
+                    ]
+                })
+            log.info(f"Found existing genai.get_file {name=}")
+            
+        except Exception as e:
+            log.info(f"Not found checking genai.get_file: '{name}' {str(e)}")
+            tasks.append(download_gcs_upload_genai(img_url, mime_type, name=name))
+
+    # Run all tasks in parallel
+    if tasks:
+        task_content = await asyncio.gather(*tasks)
+        content.extend(task_content)
+
+    return content
+
+# Helper function to handle each file download with error handling
+async def download_file_with_error_handling(img_url, mime_type, name):
+    try:
+        return await download_gcs_upload_genai(img_url, mime_type, name)
+    except Exception as err:
+        msg= f"Error processing file from {img_url}: {str(err)}"
+        log.error(msg)
+        return {"role": "user", "parts": [{"text": msg}]}
+
+async def download_gcs_upload_genai(img_url, mime_type, name=None, retries=3, delay=2):
+    import aiofiles
+    from google.generativeai.types import file_types
+    """
+    Downloads and uploads a file with retries in case of failure.
+    
+    Args:
+      - img_url: str The URL of the file to download.
+      - mime_type: str The MIME type of the file.
+      - name: str Optional name, else a random one will be created
+      - retries: int Number of retry attempts before failing.
+      - delay: int Initial delay between retries, exponentially increasing.
+      
+    Returns:
+      - downloaded_content: The result of the file upload if successful.
+    """
+    for attempt in range(retries):
+        try:
+            log.info(f"Upload {attempt} for {img_url=}")
+            # Download the file bytes asynchronously
+            file_bytes = await asyncio.to_thread(get_bytes_from_gcs, img_url)
+            if not file_bytes:
+                msg = f"Failed to download file from {img_url}: got None"
+                log.warning(msg)
+                return {"role": "user", "parts": [{"text": msg}]}
+            
+            # Log the size of the file bytes
+            file_size = len(file_bytes)
+            log.info(f"Downloaded file size for {img_url}: {file_size} bytes")
+
+            if file_size > 19434343:
+                log.warning(f"File size for {img_url}: {file_size} is too big.")
+                msg =  f"The file for {img_url} is too large ({file_size} bytes) to be used directly.  Use RAG instead."
+                return {"role": "user", "parts": [{"text": msg}]}
+            
+            extension = mimetypes.guess_extension(mime_type)
+            
+            # Use aiofiles for asynchronous file operations
+            temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=extension)
+            downloaded_file = temp_file.name
+
+            sanitized_file = sanitize_file(downloaded_file)
+
+            log.info(f"Writing file {sanitized_file}")
+            async with aiofiles.open(sanitized_file, 'wb') as f:
+                await f.write(file_bytes)
+
+            # Upload the file and get its content reference
+            try:
+                downloaded_content: file_types.File = await asyncio.to_thread(
+                    partial(genai.upload_file, name=name, mime_type=mime_type), 
+                    sanitized_file
+                    )
+                return {"role": "user", "parts": [{"file_data": downloaded_content}, 
+                                                  {"text": f"You have been given the ability to read and work with filename '{name}' with {mime_type=}."}
+                                                  ]}
+            except Exception as err:
+                msg = f"Could not upload {sanitized_file} to genai.upload_file: {str(err)} {traceback.format_exc()}"
+                log.error(msg)
+                return {"role": "user", "parts": [{"text": msg}]}
+        
+        except Exception as err:
+            log.error(f"Error processing file {img_url} on attempt {attempt + 1}/{retries}: {str(err)}")
+
+            if attempt < retries - 1:
+                log.info(f"Retrying in {delay} seconds...")
+                await asyncio.sleep(delay)
+                delay *= 2  # Exponential backoff
+            else:
+                raise err  # Raise the error after max retries
+

sunholo/senses/stream_voice.py

@@ -6,6 +6,8 @@ try:
     import sounddevice as sd
 except ImportError:
     sd = None
+except OSError:
+    sd = None
 
 try:
     import numpy as np

{sunholo-0.108.0.dist-info → sunholo-0.109.3.dist-info}/METADATA

@@ -1,9 +1,9 @@
 Metadata-Version: 2.1
 Name: sunholo
-Version: 0.108.0
+Version: 0.109.3
 Summary: Large Language Model DevOps - a package to help deploy LLMs to the Cloud.
 Home-page: https://github.com/sunholo-data/sunholo-py
-Download-URL: https://github.com/sunholo-data/sunholo-py/archive/refs/tags/v0.108.0.tar.gz
+Download-URL: https://github.com/sunholo-data/sunholo-py/archive/refs/tags/v0.109.3.tar.gz
 Author: Holosun ApS
 Author-email: multivac@sunholo.com
 License: Apache License, Version 2.0
@@ -21,8 +21,9 @@ License-File: LICENSE.txt
 Requires-Dist: google-auth
 Requires-Dist: ruamel.yaml
 Requires-Dist: langchain ==0.2.16
-Requires-Dist: langchain-experimental >=0.0.61
-Requires-Dist: langchain-community >=0.2.11
+Requires-Dist: langchain-experimental ==0.0.65
+Requires-Dist: langchain-community ==0.2.17
+Requires-Dist: langsmith ==0.1.140
 Provides-Extra: all
 Requires-Dist: anthropic[vertex] ; extra == 'all'
 Requires-Dist: asyncpg ; extra == 'all'
@@ -50,7 +51,7 @@ Requires-Dist: httpcore ; extra == 'all'
 Requires-Dist: httpx ; extra == 'all'
 Requires-Dist: jsonschema ; extra == 'all'
 Requires-Dist: lancedb ; extra == 'all'
-Requires-Dist: langchain >=0.2.12 ; extra == 'all'
+Requires-Dist: langchain >=0.2.16 ; extra == 'all'
 Requires-Dist: langchain-experimental >=0.0.61 ; extra == 'all'
 Requires-Dist: langchain-community >=0.2.11 ; extra == 'all'
 Requires-Dist: langchain-openai ==0.1.25 ; extra == 'all'

{sunholo-0.108.0.dist-info → sunholo-0.109.3.dist-info}/RECORD

@@ -85,7 +85,8 @@ sunholo/gcs/download_folder.py,sha256=ijJTnS595JqZhBH8iHFErQilMbkuKgL-bnTCMLGuvl
 sunholo/gcs/download_url.py,sha256=Ul81n1rklr8WogPsuxWWD1Nr8RHU451LzHPMJNhAKzw,6416
 sunholo/gcs/extract_and_sign.py,sha256=paRrTCvCN5vkQwCB7OSkxWi-pfOgOtZ0bwdXE08c3Ps,1546
 sunholo/gcs/metadata.py,sha256=oQLcXi4brsZ74aegWyC1JZmhlaEV270HS5_UWtAYYWE,898
-sunholo/genai/__init__.py,sha256=dBl6IA3-Fx6-Vx81r0XqxHlUq6WeW1iDX188dpChu8s,115
+sunholo/genai/__init__.py,sha256=6SWK7uV5F625J-P3xQoD6WKL59a9RSaidj-Guslyt8Q,192
+sunholo/genai/file_handling.py,sha256=DRMZlqZtZNjNmM033gvVqqxxO7J_Xvz_K78mOY5nxTQ,7496
 sunholo/genai/images.py,sha256=EyjsDqt6XQw99pZUQamomCpMOoIah9bp3XY94WPU7Ms,1678
 sunholo/genai/init.py,sha256=yG8E67TduFCTQPELo83OJuWfjwTnGZsyACospahyEaY,687
 sunholo/genai/process_funcs_cls.py,sha256=7_RQMqIAZ3nPP-GFgCHBvS39fwuWuGtvSyuJaJN_G3E,31590
@@ -116,7 +117,7 @@ sunholo/qna/__init__.py,sha256=F8q1uR_HreoSX0IfmKY1qoSwIgXhO2Q8kuDSxh9_-EE,28
 sunholo/qna/parsers.py,sha256=YpOaK5S_LxJ6FbliSYDc3AVOJ62RVduayoNnzi_p8CM,2494
 sunholo/qna/retry.py,sha256=yMw7RTkw-RXCzfENPJOt8c32mXlpvOR589EGkvK-6yI,2028
 sunholo/senses/__init__.py,sha256=fbWqVwwzkV5uRSb8lQzo4pn0ja_VYVWbUYapurSowBs,39
-sunholo/senses/stream_voice.py,sha256=tQL5TDH43OuXRlxWRLiIAceYApXyqNczN1D3jUiNMUo,18092
+sunholo/senses/stream_voice.py,sha256=VpCmooEKghBT1jPJe9mX7gKggGqY6qt-bpO7hwY4sPE,18122
 sunholo/streaming/__init__.py,sha256=MpbydI2UYo_adttPQFkxNM33b-QRyNEbrKJx0C2AGPc,241
 sunholo/streaming/content_buffer.py,sha256=0LHMwH4ctq5kjhIgMFNH0bA1RL0jMISlLVzzLcFrvv4,12766
 sunholo/streaming/langserve.py,sha256=hi7q8WY8DPKrALl9m_dOMxWOdE-iEuk7YW05SVDFIX8,6514
@@ -149,9 +150,9 @@ sunholo/vertex/init.py,sha256=1OQwcPBKZYBTDPdyU7IM4X4OmiXLdsNV30C-fee2scQ,2875
 sunholo/vertex/memory_tools.py,sha256=tBZxqVZ4InTmdBvLlOYwoSEWu4-kGquc-gxDwZCC4FA,7667
 sunholo/vertex/safety.py,sha256=S9PgQT1O_BQAkcqauWncRJaydiP8Q_Jzmu9gxYfy1VA,2482
 sunholo/vertex/type_dict_to_json.py,sha256=uTzL4o9tJRao4u-gJOFcACgWGkBOtqACmb6ihvCErL8,4694
-sunholo-0.108.0.dist-info/LICENSE.txt,sha256=SdE3QjnD3GEmqqg9EX3TM9f7WmtOzqS1KJve8rhbYmU,11345
-sunholo-0.108.0.dist-info/METADATA,sha256=NoRyTCRajNC5x0uAVpyUs8Ogp9oCuChGi2rcMdpG5QQ,8670
-sunholo-0.108.0.dist-info/WHEEL,sha256=P9jw-gEje8ByB7_hXoICnHtVCrEwMQh-630tKvQWehc,91
-sunholo-0.108.0.dist-info/entry_points.txt,sha256=bZuN5AIHingMPt4Ro1b_T-FnQvZ3teBes-3OyO0asl4,49
-sunholo-0.108.0.dist-info/top_level.txt,sha256=wt5tadn5--5JrZsjJz2LceoUvcrIvxjHJe-RxuudxAk,8
-sunholo-0.108.0.dist-info/RECORD,,
+sunholo-0.109.3.dist-info/LICENSE.txt,sha256=SdE3QjnD3GEmqqg9EX3TM9f7WmtOzqS1KJve8rhbYmU,11345
+sunholo-0.109.3.dist-info/METADATA,sha256=kcRA_wA-X5fPF1zpAU-nyDdHmTVgwUHpDQgqeYcmQH0,8705
+sunholo-0.109.3.dist-info/WHEEL,sha256=P9jw-gEje8ByB7_hXoICnHtVCrEwMQh-630tKvQWehc,91
+sunholo-0.109.3.dist-info/entry_points.txt,sha256=bZuN5AIHingMPt4Ro1b_T-FnQvZ3teBes-3OyO0asl4,49
+sunholo-0.109.3.dist-info/top_level.txt,sha256=wt5tadn5--5JrZsjJz2LceoUvcrIvxjHJe-RxuudxAk,8
+sunholo-0.109.3.dist-info/RECORD,,