ai-parrot 0.3.3__tar.gz → 0.3.6__tar.gz

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/.github/workflows/release.yml

@@ -26,7 +26,7 @@ jobs:
 
       - name: Build wheels on Ubuntu
         if: matrix.os == 'ubuntu-latest'
-        uses: RalfG/python-wheels-manylinux-build@v0.5.0-manylinux2014_x86_64
+        uses: RalfG/python-wheels-manylinux-build@v0.7.1-manylinux2014_x86_64
         with:
           python-versions: 'cp39-cp39 cp310-cp310 cp311-cp311 cp312-cp312'
           build-requirements: 'cython numpy'

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/.gitignore

@@ -175,3 +175,6 @@ docs/videos/*.mkv
 docs/videos/*.mp3
 docs/vimeo/*
 docs/youtube/*
+
+# more documents:
+documents/*

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/Makefile

@@ -14,7 +14,7 @@ install:
 	# QS requirements
 	pip install --upgrade querysource[analytics]
 	# and Parrot:
-	# pip install -e .[google,milvus,analytics]
+	# pip install -e .[google,milvus,groq,analytics]
 
 develop:
 	# Install Parrot

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ai-parrot
-Version: 0.3.3
+Version: 0.3.6
 Summary: Live Chatbots based on Langchain chatbots and Agents     Integrated into Navigator Framework or used into aiohttp applications.
 Home-page: https://github.com/phenobarbital/ai-parrot
 Author: Jesus Lara
@@ -26,7 +26,7 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Framework :: AsyncIO
-Requires-Python: >=3.10.12
+Requires-Python: >=3.9.20
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: Cython==3.0.11
@@ -58,11 +58,8 @@ Requires-Dist: sentence-transformers==3.0.1
 Requires-Dist: tabulate==0.9.0
 Requires-Dist: tiktoken==0.7.0
 Requires-Dist: tokenizers==0.19.1
-Requires-Dist: unstructured==0.14.3
-Requires-Dist: unstructured-client==0.18.0
-Requires-Dist: youtube-transcript-api==0.6.2
-Requires-Dist: selenium==4.18.1
-Requires-Dist: webdriver_manager==4.0.1
+Requires-Dist: selenium>=4.18.1
+Requires-Dist: webdriver_manager>=4.0.1
 Requires-Dist: transitions==0.9.0
 Requires-Dist: sentencepiece==0.2.0
 Requires-Dist: duckduckgo-search==5.3.0
@@ -79,9 +76,12 @@ Requires-Dist: mediawikiapi==1.2
 Requires-Dist: pyowm==3.3.0
 Requires-Dist: O365==2.0.35
 Requires-Dist: stackapi==0.3.1
-Requires-Dist: timm==1.0.9
 Requires-Dist: torchvision==0.19.1
+Requires-Dist: tf-keras==2.17.0
 Provides-Extra: loaders
+Requires-Dist: unstructured==0.14.3; extra == "loaders"
+Requires-Dist: unstructured-client==0.18.0; extra == "loaders"
+Requires-Dist: youtube-transcript-api==0.6.2; extra == "loaders"
 Requires-Dist: pymupdf==1.24.4; extra == "loaders"
 Requires-Dist: pymupdf4llm==0.0.1; extra == "loaders"
 Requires-Dist: pdf4llm==0.0.6; extra == "loaders"
@@ -107,7 +107,7 @@ Requires-Dist: paddleocr==2.8.1; extra == "loaders"
 Requires-Dist: ftfy==6.2.3; extra == "loaders"
 Requires-Dist: librosa==0.10.1; extra == "loaders"
 Requires-Dist: XlsxWriter==3.2.0; extra == "loaders"
-Requires-Dist: xformers==0.0.27.post2; extra == "loaders"
+Requires-Dist: timm==1.0.9; extra == "loaders"
 Provides-Extra: anthropic
 Requires-Dist: langchain-anthropic==0.1.11; extra == "anthropic"
 Requires-Dist: anthropic==0.25.2; extra == "anthropic"
@@ -142,24 +142,6 @@ Requires-Dist: gradio-client==0.2.9; extra == "analytics"
 Requires-Dist: streamlit==1.37.1; extra == "analytics"
 Requires-Dist: simsimd==4.3.1; extra == "analytics"
 Requires-Dist: opencv-python==4.10.0.84; extra == "analytics"
-Provides-Extra: all
-Requires-Dist: langchain-milvus==0.1.1; extra == "all"
-Requires-Dist: milvus==2.3.5; extra == "all"
-Requires-Dist: pymilvus==2.4.4; extra == "all"
-Requires-Dist: groq==0.11.0; extra == "all"
-Requires-Dist: langchain-groq==0.1.4; extra == "all"
-Requires-Dist: llama-index-llms-huggingface==0.2.7; extra == "all"
-Requires-Dist: langchain-google-vertexai==1.0.8; extra == "all"
-Requires-Dist: langchain-google-genai==1.0.8; extra == "all"
-Requires-Dist: google-generativeai==0.7.2; extra == "all"
-Requires-Dist: vertexai==1.60.0; extra == "all"
-Requires-Dist: google-cloud-aiplatform>=1.60.0; extra == "all"
-Requires-Dist: grpc-google-iam-v1==0.13.0; extra == "all"
-Requires-Dist: langchain-openai==0.1.21; extra == "all"
-Requires-Dist: openai==1.40.8; extra == "all"
-Requires-Dist: llama-index-llms-openai==0.1.11; extra == "all"
-Requires-Dist: langchain-anthropic==0.1.23; extra == "all"
-Requires-Dist: anthropic==0.34.0; extra == "all"
 
 # AI Parrot: Python package for creating Chatbots
 This is an open-source Python package for creating Chatbots based on Langchain and Navigator.

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/ai_parrot.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ai-parrot
-Version: 0.3.3
+Version: 0.3.6
 Summary: Live Chatbots based on Langchain chatbots and Agents     Integrated into Navigator Framework or used into aiohttp applications.
 Home-page: https://github.com/phenobarbital/ai-parrot
 Author: Jesus Lara
@@ -26,7 +26,7 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Framework :: AsyncIO
-Requires-Python: >=3.10.12
+Requires-Python: >=3.9.20
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: Cython==3.0.11
@@ -58,11 +58,8 @@ Requires-Dist: sentence-transformers==3.0.1
 Requires-Dist: tabulate==0.9.0
 Requires-Dist: tiktoken==0.7.0
 Requires-Dist: tokenizers==0.19.1
-Requires-Dist: unstructured==0.14.3
-Requires-Dist: unstructured-client==0.18.0
-Requires-Dist: youtube-transcript-api==0.6.2
-Requires-Dist: selenium==4.18.1
-Requires-Dist: webdriver_manager==4.0.1
+Requires-Dist: selenium>=4.18.1
+Requires-Dist: webdriver_manager>=4.0.1
 Requires-Dist: transitions==0.9.0
 Requires-Dist: sentencepiece==0.2.0
 Requires-Dist: duckduckgo-search==5.3.0
@@ -79,9 +76,12 @@ Requires-Dist: mediawikiapi==1.2
 Requires-Dist: pyowm==3.3.0
 Requires-Dist: O365==2.0.35
 Requires-Dist: stackapi==0.3.1
-Requires-Dist: timm==1.0.9
 Requires-Dist: torchvision==0.19.1
+Requires-Dist: tf-keras==2.17.0
 Provides-Extra: loaders
+Requires-Dist: unstructured==0.14.3; extra == "loaders"
+Requires-Dist: unstructured-client==0.18.0; extra == "loaders"
+Requires-Dist: youtube-transcript-api==0.6.2; extra == "loaders"
 Requires-Dist: pymupdf==1.24.4; extra == "loaders"
 Requires-Dist: pymupdf4llm==0.0.1; extra == "loaders"
 Requires-Dist: pdf4llm==0.0.6; extra == "loaders"
@@ -107,7 +107,7 @@ Requires-Dist: paddleocr==2.8.1; extra == "loaders"
 Requires-Dist: ftfy==6.2.3; extra == "loaders"
 Requires-Dist: librosa==0.10.1; extra == "loaders"
 Requires-Dist: XlsxWriter==3.2.0; extra == "loaders"
-Requires-Dist: xformers==0.0.27.post2; extra == "loaders"
+Requires-Dist: timm==1.0.9; extra == "loaders"
 Provides-Extra: anthropic
 Requires-Dist: langchain-anthropic==0.1.11; extra == "anthropic"
 Requires-Dist: anthropic==0.25.2; extra == "anthropic"
@@ -142,24 +142,6 @@ Requires-Dist: gradio-client==0.2.9; extra == "analytics"
 Requires-Dist: streamlit==1.37.1; extra == "analytics"
 Requires-Dist: simsimd==4.3.1; extra == "analytics"
 Requires-Dist: opencv-python==4.10.0.84; extra == "analytics"
-Provides-Extra: all
-Requires-Dist: langchain-milvus==0.1.1; extra == "all"
-Requires-Dist: milvus==2.3.5; extra == "all"
-Requires-Dist: pymilvus==2.4.4; extra == "all"
-Requires-Dist: groq==0.11.0; extra == "all"
-Requires-Dist: langchain-groq==0.1.4; extra == "all"
-Requires-Dist: llama-index-llms-huggingface==0.2.7; extra == "all"
-Requires-Dist: langchain-google-vertexai==1.0.8; extra == "all"
-Requires-Dist: langchain-google-genai==1.0.8; extra == "all"
-Requires-Dist: google-generativeai==0.7.2; extra == "all"
-Requires-Dist: vertexai==1.60.0; extra == "all"
-Requires-Dist: google-cloud-aiplatform>=1.60.0; extra == "all"
-Requires-Dist: grpc-google-iam-v1==0.13.0; extra == "all"
-Requires-Dist: langchain-openai==0.1.21; extra == "all"
-Requires-Dist: openai==1.40.8; extra == "all"
-Requires-Dist: llama-index-llms-openai==0.1.11; extra == "all"
-Requires-Dist: langchain-anthropic==0.1.23; extra == "all"
-Requires-Dist: anthropic==0.34.0; extra == "all"
 
 # AI Parrot: Python package for creating Chatbots
 This is an open-source Python package for creating Chatbots based on Langchain and Navigator.

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/ai_parrot.egg-info/SOURCES.txt

@@ -22,7 +22,12 @@ ai_parrot.egg-info/SOURCES.txt
 ai_parrot.egg-info/dependency_links.txt
 ai_parrot.egg-info/requires.txt
 ai_parrot.egg-info/top_level.txt
+documents/AR_Certification_Skill_Practice_Scorecard_EXAMPLE.pdf
+documents/Day 1_Essentials_AR_PPT.pdf
 documents/ex-code-loaders.txt
+documents/video_2024-09-11_19-43-58.mp3
+documents/video_2024-09-11_19-43-58.mp4
+documents/video_2024-09-11_19-43-58.vtt
 etc/navigator-ssl.ini
 etc/navigator.ini
 etc/ssl/domain.ext
@@ -32,7 +37,9 @@ etc/ssl/navigator.local.key
 etc/ssl/rootCA.crt
 etc/ssl/rootCA.key
 etc/ssl/rootCA.srl
+examples/analyze_video.py
 examples/check_bot.py
+examples/extract_frames.py
 examples/load_pdf.py
 examples/test_bot.py
 examples/test_question.py

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/ai_parrot.egg-info/requires.txt

@@ -27,11 +27,8 @@ sentence-transformers==3.0.1
 tabulate==0.9.0
 tiktoken==0.7.0
 tokenizers==0.19.1
-unstructured==0.14.3
-unstructured-client==0.18.0
-youtube-transcript-api==0.6.2
-selenium==4.18.1
-webdriver_manager==4.0.1
+selenium>=4.18.1
+webdriver_manager>=4.0.1
 transitions==0.9.0
 sentencepiece==0.2.0
 duckduckgo-search==5.3.0
@@ -48,27 +45,8 @@ mediawikiapi==1.2
 pyowm==3.3.0
 O365==2.0.35
 stackapi==0.3.1
-timm==1.0.9
 torchvision==0.19.1
-
-[all]
-langchain-milvus==0.1.1
-milvus==2.3.5
-pymilvus==2.4.4
-groq==0.11.0
-langchain-groq==0.1.4
-llama-index-llms-huggingface==0.2.7
-langchain-google-vertexai==1.0.8
-langchain-google-genai==1.0.8
-google-generativeai==0.7.2
-vertexai==1.60.0
-google-cloud-aiplatform>=1.60.0
-grpc-google-iam-v1==0.13.0
-langchain-openai==0.1.21
-openai==1.40.8
-llama-index-llms-openai==0.1.11
-langchain-anthropic==0.1.23
-anthropic==0.34.0
+tf-keras==2.17.0
 
 [analytics]
 annoy==1.17.3
@@ -100,6 +78,9 @@ langchain-groq==0.1.9
 llama-index-llms-huggingface==0.2.7
 
 [loaders]
+unstructured==0.14.3
+unstructured-client==0.18.0
+youtube-transcript-api==0.6.2
 pymupdf==1.24.4
 pymupdf4llm==0.0.1
 pdf4llm==0.0.6
@@ -125,7 +106,7 @@ paddleocr==2.8.1
 ftfy==6.2.3
 librosa==0.10.1
 XlsxWriter==3.2.0
-xformers==0.0.27.post2
+timm==1.0.9
 
 [milvus]
 langchain-milvus>=0.1.4

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/app.py

@@ -3,7 +3,7 @@ from navigator.handlers.types import AppHandler
 from navigator.background import BackgroundQueue
 from navigator_auth import AuthHandler
 from parrot.manager import ChatbotManager
-from parrot.loaders.handlers import DataManagement
+# from parrot.loaders.handlers import DataManagement
 from parrot.conf import STATIC_DIR
 from parrot.handlers.bots import (
     FeedbackTypeHandler,
@@ -53,7 +53,7 @@ class Main(AppHandler):
             ChatbotSharingQuestion
         )
         # Management APIs:
-        DataManagement.configure(self.app)
+        # DataManagement.configure(self.app)
 
 
     async def on_prepare(self, request, response):

ai_parrot-0.3.6/documents/video_2024-09-11_19-43-58.vtt

@@ -0,0 +1,122 @@
+WEBVTT
+
+1
+00:00:00.000 --> 00:00:09.320
+ Hi, I'm Kyle from the Bose Support Team and here on the Bose Team we take merchandising
+
+2
+00:00:09.320 --> 00:00:11.980
+ and brand advocacy to the next level.
+
+3
+00:00:11.980 --> 00:00:16.540
+ Every visit starts with arriving at the location and greeting an associate by the door.
+
+4
+00:00:16.540 --> 00:00:19.000
+ We move on to the merchandising side of our visit.
+
+5
+00:00:19.000 --> 00:00:23.440
+ We head over to our displays to ensure that they're clean, bright and fully stocked.
+
+6
+00:00:23.440 --> 00:00:27.800
+ After our displays are clean and bright, we need to test the functionality of these displays. We go
+
+7
+00:00:27.800 --> 00:00:32.220
+ through each display ensuring that the demo is functioning as intended. If we
+
+8
+00:00:32.220 --> 00:00:36.980
+ run into an issue, we troubleshoot immediately or we call the call center
+
+9
+00:00:36.980 --> 00:00:41.300
+ for further support. This can be more in-depth troubleshooting or to order
+
+10
+00:00:41.300 --> 00:00:48.400
+ parts that are needed to resolve the issue. These can be simple fixes such as installing a wire, swapping out the media card, or even
+
+11
+00:00:48.400 --> 00:00:49.400
+ a simple reboot.
+
+12
+00:00:49.400 --> 00:00:54.680
+ They can even be more advanced fixes such as entire display rewiring, product changeouts,
+
+13
+00:00:54.680 --> 00:00:56.920
+ or even full display changeouts.
+
+14
+00:00:56.920 --> 00:01:01.280
+ Our ticketing system houses all of our tickets and allows for reps to properly manage their
+
+15
+00:01:01.280 --> 00:01:05.000
+ market opening, closing tickets, as well as seeing when parts
+
+16
+00:01:05.000 --> 00:01:06.280
+ have been shipped or delivered.
+
+17
+00:01:06.280 --> 00:01:11.120
+ Once we've ensured that our displays are properly merchandised and fully functional,
+
+18
+00:01:11.120 --> 00:01:13.720
+ we move on to the brand advocacy part of our visit.
+
+19
+00:01:13.720 --> 00:01:17.520
+ We do this in several ways, by building the relationship with the store, this is done
+
+20
+00:01:17.520 --> 00:01:22.700
+ through routine check-ins with staff and management in the field to cover common concerns or questions
+
+21
+00:01:22.700 --> 00:01:28.160
+ that they may have, as well as fill them in on some new features that they may not know. Training events such
+
+22
+00:01:28.160 --> 00:01:32.880
+ as morning huddles, lunch and learns, or customized training experiences where we
+
+23
+00:01:32.880 --> 00:01:37.360
+ can take a full team sit down and do some in-depth training. Demoing new
+
+24
+00:01:37.360 --> 00:01:42.040
+ products letting both customers and store associates get a hands-on
+
+25
+00:01:42.040 --> 00:01:49.640
+ experience with a newly released product and promotion events where we have an opportunity to get on the front line and practice what we preach,
+
+26
+00:01:49.640 --> 00:01:51.780
+ demoing and selling them ourselves.
+
+27
+00:01:51.780 --> 00:01:57.680
+ Through these regular engagements, we are able to impact not just a store but an entire
+
+28
+00:01:57.680 --> 00:01:58.680
+ market.
+
+29
+00:01:58.680 --> 00:02:02.960
+ This is just a quick walkthrough of what we do day to day here at the Bose program.
+
+30
+00:02:02.960 --> 00:02:05.000
+ I hope you enjoyed and have a great day.
+

ai_parrot-0.3.6/examples/analyze_video.py

@@ -0,0 +1,33 @@
+import asyncio
+from navconfig import BASE_DIR
+from parrot.llms.vertex import VertexLLM
+from parrot.loaders.videolocal import (
+    VideoLocalLoader
+)
+
+
+
+
+async def process_video(doc):
+    llm = VertexLLM(
+        model='gemini-1.5-pro',
+        temperature=0.1,
+        top_k=30,
+        Top_p=0.5,
+    )
+    print(':: Processing: ', doc)
+    loader = VideoLocalLoader(
+        doc,
+        source_type=f"Video {doc.name}",
+        llm=llm.get_llm(),
+        language="en"
+    )
+    docs = loader.extract()
+    print('DOCS > ', docs)
+
+
+if __name__ == '__main__':
+    doc = BASE_DIR.joinpath('documents', 'video_2024-09-11_19-43-58.mp4')
+    asyncio.run(
+        process_video(doc)
+    )

ai_parrot-0.3.6/examples/extract_frames.py

@@ -0,0 +1,74 @@
+import os
+from pathlib import PurePath, Path
+import cv2
+from navconfig import BASE_DIR
+
+def better_resolution(image_path, output_path):
+    # Load the super-resolution model (example using the EDSR model)
+    sr = cv2.dnn_superres.DnnSuperResImpl_create()
+
+    # Read the model
+    path_to_model = 'EDSR_x4.pb'
+    # You need to download this model from OpenCV's model zoo
+    sr.readModel(path_to_model)
+
+    # Set the model and scale
+    sr.setModel("edsr", 4)  # EDSR model with 4x upscaling
+
+    # Read the input image
+    image = cv2.imread(image_path)
+
+    # Upscale the image
+    upscaled_image = sr.upsample(image)
+
+    # Save the result
+    cv2.imwrite(output_path, upscaled_image)
+    print(f"Saved super-resolution image: {output_path}")
+
+def extract_frames(
+    video_path,
+    output_dir: PurePath,
+    interval=5,
+    upscale_factor=2
+):
+    if not output_dir.exists():
+        output_dir.mkdir(mode=0o777, parents=True, exist_ok=True)
+
+    cap = cv2.VideoCapture(str(video_path))
+
+    # Get frames per second (fps) of the video
+    fps = cap.get(cv2.CAP_PROP_FPS)
+    frame_interval = int(fps * interval)
+
+    frame_count = 0
+    success, frame = cap.read()
+
+    while success:
+        if frame_count % frame_interval == 0:
+            # Get the original dimensions
+            height, width = frame.shape[:2]
+            # Upscale the frame by the given factor
+            frame_upscaled = cv2.resize(
+                frame,
+                (width * upscale_factor, height * upscale_factor),
+                interpolation=cv2.INTER_CUBIC
+            )
+            frame_name = f"frame_{frame_count}.jpg"
+            upscaled_name = f"frame_{frame_count}_upscaled.jpg"
+            frame_path = os.path.join(output_dir, frame_name)
+            upscaled_path = output_dir.joinpath(upscaled_name)
+            cv2.imwrite(frame_path, frame_upscaled)
+            # better_resolution(frame_path, upscaled_path)
+            print(f"Extracted {frame_name}")
+
+        frame_count += 1
+        success, frame = cap.read()
+
+    cap.release()
+    print("Finished extracting frames.")
+
+# Usage
+if __name__ == '__main__':
+    video_file = BASE_DIR.joinpath('documents', 'video_2024-09-11_19-43-58.mp4')
+    output_folder = BASE_DIR.joinpath('documents', 'extracted_frames')
+    extract_frames(video_file, output_folder, interval=2)

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/examples/load_pdf.py

@@ -7,13 +7,14 @@ from parrot.loaders import (
 
 async def process_pdf():
     llm = VertexLLM(
-        model='gemini-1.5-flash-001',
+        model='gemini-1.5-pro',
         temperature=0.1,
         top_k=30,
         Top_p=0.5,
     )
     # Add LLM
-    doc = BASE_DIR.joinpath('documents', 'AR_Certification_Skill_Practice_Scorecard_EXAMPLE.pdf')
+    # doc = BASE_DIR.joinpath('documents', 'AR_Certification_Skill_Practice_Scorecard_EXAMPLE.pdf')
+    doc = BASE_DIR.joinpath('documents', 'Day 1_Essentials_AR_PPT.pdf')
     print(':: Processing: ', doc)
     # PDF Files
     loader = PDFLoader(

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/examples/test_bot.py

@@ -13,7 +13,7 @@ async def get_agent():
     #     Top_p=0.6,
     # )
     llm = VertexLLM(
-        model='gemini-1.5-flash-001',
+        model='gemini-1.5-pro',
         temperature=0.1,
         top_k=30,
         Top_p=0.5,

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/parrot/loaders/pdf.py

@@ -49,14 +49,6 @@ class PDFLoader(BasePDF):
         self.parse_images = kwargs.get('parse_images', False)
         self.page_as_images = kwargs.get('page_as_images', False)
         if self.page_as_images is True:
-            # # Load the processor and model from Hugging Face
-            # self.image_processor = DonutProcessor.from_pretrained(
-            #     "naver-clova-ix/donut-base-finetuned-docvqa"
-            # )
-            # self.image_model = VisionEncoderDecoderModel.from_pretrained(
-            #     "naver-clova-ix/donut-base-finetuned-docvqa",
-
-            # )
             # Load the processor and model from Hugging Face
             self.image_processor = LayoutLMv3Processor.from_pretrained(
                 "microsoft/layoutlmv3-base",
@@ -388,6 +380,7 @@ class PDFLoader(BasePDF):
                     # TODO passing the image to a AI visual to get explanation
                     # Get the extracted text from the image
                     text = self.extract_page_text(img_path)
+                    print('TEXT EXTRACTED >> ', text)
                     url = f'/static/images/{img_name}'
                     image_meta = {
                         "url": url,

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/parrot/loaders/videolocal.py

@@ -1,10 +1,26 @@
 from typing import Any
 from collections.abc import Callable
+import math
 from pathlib import PurePath
 from langchain.docstore.document import Document
 from .basevideo import BaseVideoLoader
 
 
+def split_text(text, max_length):
+    """Split text into chunks of a maximum length, ensuring not to break words."""
+    chunks = []
+    while len(text) > max_length:
+        # Find the last space before the max_length
+        split_point = text.rfind(' ', 0, max_length)
+        # If no space found, split at max_length
+        if split_point == -1:
+            split_point = max_length
+        chunks.append(text[:split_point])
+        text = text[split_point:].strip()
+    chunks.append(text)
+    return chunks
+
+
 class VideoLocalLoader(BaseVideoLoader):
     """
     Generating Video transcripts from local Videos.
@@ -26,14 +42,15 @@ class VideoLocalLoader(BaseVideoLoader):
 
     def load_video(self, path: PurePath) -> list:
         metadata = {
-            "source": f"{path}",
             "url": f"{path.name}",
-            "index": path.stem,
+            "source": f"{path}",
             "filename": f"{path}",
+            "index": path.stem,
             "question": '',
             "answer": '',
             'type': 'video_transcript',
             "source_type": self._source_type,
+            "data": {},
             "summary": '',
             "document_meta": {
                 "language": self._language,
@@ -53,14 +70,17 @@ class VideoLocalLoader(BaseVideoLoader):
             transcript = ''
         # Summarize the transcript
         if transcript:
+            # Split transcript into chunks
+            transcript_chunks = split_text(transcript, 32767)
             summary = self.get_summary_from_text(transcript)
             # Create Two Documents, one is for transcript, second is VTT:
             metadata['summary'] = summary
-            doc = Document(
-                page_content=transcript,
-                metadata=metadata
-            )
-            documents.append(doc)
+            for chunk in transcript_chunks:
+                doc = Document(
+                    page_content=chunk,
+                    metadata=metadata
+                )
+                documents.append(doc)
         if transcript_whisper:
             # VTT version:
             transcript = self.transcript_to_vtt(transcript_whisper, transcript_path)

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/parrot/version.py

@@ -3,7 +3,7 @@
 __title__ = "ai-parrot"
 __description__ = "Live Chatbots based on Langchain chatbots and Agents \
     Integrated into Navigator Framework or used into aiohttp applications."
-__version__ = "0.3.3"
+__version__ = "0.3.6"
 __author__ = "Jesus Lara"
 __author_email__ = "jesuslarag@gmail.com"
 __license__ = "MIT"

{ai_parrot-0.3.3 → ai_parrot-0.3.6}/pyproject.toml

@@ -34,7 +34,7 @@ classifiers=[
   "License :: OSI Approved :: MIT License",
 ]
 description-file = "README.md"
-requires-python = ">=3.10.12"
+requires-python = ">=3.9.20"
 
 [tool.pytest.ini_options]
 addopts = [