Enhancing Document-Centric Features with On-Premise Generative AI for Alfresco Community
0 likes146 views
Oractical guide on integrating Alfresco Community with On-Premise Generative AI. This session outlines the steps to enhance both existing and new content, demonstrating features such as classification, summarization, translation, and prompting. But this framework allows you to include additional features. Source code is available in https://github.com/aborroy/alfresco-genai
![Document Features
New Content
alfresco-ai-listener REST API
ACTIVEMQ
Listen to ASPECT setting
Apache License 2.0
https://github.com/aborroy/alfresco-genai
document.move(
document.parent.childByNamePath(
document.properties["genai:term"]
)
);
classifiable
terms text, multiple
summarizable
summary text
tags text
llmSummary text
promptable
question text
answer text
llmPrompt text
classified
term text
llmClassify text](https://image.slidesharecdn.com/alfresco-ttl-154-genai-240124113309-439c0def/85/Enhancing-Document-Centric-Features-with-On-Premise-Generative-AI-for-Alfresco-Community-24-320.jpg)
More Related Content
What's hot (20)
Similar to Enhancing Document-Centric Features with On-Premise Generative AI for Alfresco Community (20)
![Learning Your Way Around Alfresco [A Developer's Intro, Part 1. by Jeff Potts]](https://cdn.slidesharecdn.com/ss_thumbnails/intro-to-alfresco-part-1-090414034555-phpapp01-thumbnail.jpg?width=560&fit=bounds)
More from Angel Borroy López (20)
Recently uploaded (20)
Enhancing Document-Centric Features with On-Premise Generative AI for Alfresco Community
- 2. January 19, 2024 January 19, 2024 ©2024 Hyland Software, Inc. and its affiliates. All rights reserved. All Hyland product names are registered or unregistered trademarks of Hyland Software, Inc. or its affiliates in the United States and other countries. Alfresco Tech Talk Live #154 Enhancing Document-Centric Features with On- Premise Generative AI for Alfresco Community Angel Borroy Developer Evangelist
- 3. Agenda Content credentials Generated with AI ∙ 19 January 2024 at 8:24 am
- 5. Docker GenAI Stack Components ollama • Local management of open source LLMs • Catalog of preconfigured LLMs, such as Llama2 or Mistral neo4j • Graph and native vector search capabilities • Ground LLMs for more precise GenAI predictions and outcomes LangChain • Communication between the LLM, your application, and the database • Python framework for developing applications powered by LLMs https://github.com/docker/genai-stack
- 6. Docker GenAI Deployment https://github.com/docker/genai-stack pull-model ollama neo4j streamlit LangChain REST API GPU
- 7. 2 GB Size Docker GenAI https://ollama.ai/library LLM 3 Billion Parameters phi MIT License dolphin-phi MIT License orca-mini cc-by-nc-sa-4.0 deepseek-coder DEEPSEEK LICENSE AGREEMENT 4 GB RAM
- 8. Docker GenAI https://ollama.ai/library LLM 7 Billion Parameters llama2 LLAMA2 COMMUNITY LICENSE AGREEMENT codellama LLAMA2 COMMUNITY LICENSE AGREEMENT vicuna LLAMA2 COMMUNITY LICENSE AGREEMENT mistral Apache License 2.0 mistral-openorca Apache License 2.0 llava Apache License 2.0 orca-mini cc-by-nc-sa-4.0 deepseek-coder DEEPSEEK LICENSE AGREEMENT orca2 MICROSOFT RESEARCH LICENSE falcon FALCON 180B TII LICENSE 1.0 4 GB Size 8 GB RAM
- 9. Docker GenAI https://ollama.ai/library LLM 13 Billion Parameters llama2 LLAMA2 COMMUNITY LICENSE AGREEMENT codellama LLAMA2 COMMUNITY LICENSE AGREEMENT vicuna LLAMA2 COMMUNITY LICENSE AGREEMENT orca-mini cc-by-nc-sa-4.0 llava Apache License 2.0 orca2 MICROSOFT RESEARCH LICENSE 8 GB Size 16 GB RAM
- 10. Docker GenAI https://ollama.ai/library LLM 13+ Billion Parameters llama2 LLAMA2 COMMUNITY LICENSE AGREEMENT 70B codellama LLAMA2 COMMUNITY LICENSE AGREEMENT 34B mixtral Apache License 2.0 56B deepseek-coder DEEPSEEK LICENSE AGREEMENT 33B vicuna LLAMA2 COMMUNITY LICENSE AGREEMENT 33B falcon FALCON 180B TII LICENSE 1.0 40B / 180B 40+ GB Size 64+ GB RAM
- 11. Docker GenAI Creating a new REST API streamlit langchain.chains langchain.vectorstores.neo4j_vector fastapi
- 13. Alfresco Integration Deployment https://github.com/aborroy/alfresco-genai ALFRESCO GENAI STACK Docker Network model-repo model-share llm neo4j langchain compose compose alfresco-ai-applier alfresco-ai-listener app service http://genai:8506/summary http://genai:8506/classify http://genai:8506/prompt http://alfresco:8080 tcp://activemq:61616
- 14. Alfresco Integration Alfresco Aspects model-repo model-share summarizable summary text tags text llmSummary text classifiable terms text, multiple classified term text llmClassify text promptable question text answer text llmPrompt text
- 15. Alfresco Integration AI Applier Run action on a Repository Folder • Summary • Classify • List of terms required as parameter alfresco-ai-applier ALFRESCO GENAI STACK compose compose alfresco-ai-applier app http://genai:8506/summary http://genai:8506/classify http://alfresco:8080 $ java -jar alfresco-ai-applier-0.8.0.jar --applier.root.folder=/app:company_home/app:shared --applier.action=SUMMARY $ java -jar alfresco-ai-applier-0.8.0.jar --applier.root.folder=/app:company_home/app:shared --applier.action.classify.term.list=English,Spanish --applier.action=CLASSIFY
- 16. Alfresco Integration AI Listener Listen to ActiveMQ for aspect settings • Summary • Classification • Prompting alfresco-ai-listener ALFRESCO GENAI STACK compose compose alfresco-ai-listener service http://genai:8506/summary http://genai:8506/classify http://genai:8506/prompt http://alfresco:8080 tcp://activemq:61616 alfresco-ai-listener: image: alfresco-ai-listener environment: CONTENT_SERVICE_URL: "http://alfresco:8080" SPRING_ACTIVEMQ_BROKERURL: "tcp://activemq:61616" GENAI_URL: "http://genai:8506"
- 17. GENAI STACK Alfresco Integration GenAI App Python application that provides REST API endpoints Document Loader Split document into smaller chunks LLM Embeddings Store in VectorDB Embed and save chunks Create QA Chain With the ability to lookup Vector DB LLM LLM=mistral SUMMARY_LANGUAGE=English SUMMARY_SIZE=120 TAGS_NUMBER=3
- 18. GENAI STACK Alfresco Integration GenAI App curl --location 'http://localhost:8506/summary' --form 'file=@"./file.pdf"' { "summary": " The text discusses...", "tags": " Golang, Merkle, Difficulty", "model": "mistral" } SUMMARY Write a short summary of the text in 120 words only in ENGLISH TAGS Provide 3 words to categorize the document in language ENGLISH in a single line. Use only language ENGLISH for these 3 words in the answer. Don't add any explanation for the words in the answer. Don't add any note after the list of words in the answer. Don't use bullets or numbers to list the words in the answer. Don't add in the answer the translation of the words in a different language after the list of words. Give the answer exactly as a list of 3 words in language ENGLISH separated with comma and without ending dot.
- 19. GENAI STACK Alfresco Integration GenAI App curl --location 'http://localhost:8506/classify?termList="Japanese,Spanish,Vietnamese"' --form 'file=./file.pdf"' { "term": " English", "model": "mistral" } TERM Pick one of the following list of categories: Japanese, Spanish, Vietnamese. Write the answer only in ENGLISH language. Don't add any explanation for the choice in the answer. Don't add any note after the word in the answer. Don't add any space before the word in the answer. Don't add in the answer the translation of the word in a different language after chosen word. Give the answer exactly as a single word from the list.
- 20. GENAI STACK Alfresco Integration GenAI App curl --location 'http://localhost:8506/prompt?prompt="What is this text about?”' --form 'file=./file.pdf"' { "answer": " Yes, it is difficult to find childcare in Tokyo.", "model": "mistral" } PROMPT What is this text about? Write the answer only in ENGLISH language. Don't add any translation to the answer.
- 22. Document Features Existing Content alfresco-ai-applier REST API REST API Scan a SOURCE FOLDER Apache License 2.0 https://github.com/aborroy/alfresco-genai summarizable summary text tags text llmSummary text classified term text llmClassify text
- 23. Demo Time c5a.8xlarge 32 vCPUs 64 GiB RAM
- 24. Document Features New Content alfresco-ai-listener REST API ACTIVEMQ Listen to ASPECT setting Apache License 2.0 https://github.com/aborroy/alfresco-genai document.move( document.parent.childByNamePath( document.properties["genai:term"] ) ); classifiable terms text, multiple summarizable summary text tags text llmSummary text promptable question text answer text llmPrompt text classified term text llmClassify text
- 25. Demo Time c5a.8xlarge 32 vCPUs 64 GiB RAM
- 26. What Else?
- 27. Find your way Docker AI/ML Hack-a-thon • Readme AI • Techdocs • Docker Image Analyzer • Docker Log Sentiment Analyzer • GitChats AI ollama alternatives • https://gpt4all.io/index.html • https://www.secondstate.io/run-llm • https://huggingface.co/docs/hub/spaces-sdks-docker-first-demo
- 28. Thanks!