New version (#56)

* adding pre-puller for pipeline images (#139)

* adding pre-puller for pipeline images

* Predicting the same values

* requested changes for section 5

* quick improvements.

* cleaning trailing dots

* spelling

* link to contributors

* Lowered the requirement for acceptable response similarity

* Removing model_sha as it seems volatile

* Updated 03/06 to use VLLM and improved response time

Removed streaming and fixed the response quality

Updated requirements for VLLM

Updated responsetime to hit the text generation endpoint

* minor fixes in lab1

* minor fixes in lab2 and update images from 2.8

* minor fixes in lab4 and update images for serving in 2.8

* minor adjustments in lab6

* Update to vLLM and Ollama for model serving

* Better leverage antora variables (#170)

* add antora rhoai, argocd and ocp vars

* added lab vars

* minor typo fix

* update theme for Summit (#171)

* Started drafting the instructions for DIY vs AUTO on the projects (#176)

* 02-02 diy vs auto - start

* split up of 02-03

* disclaimer at top of diy versions

* wordlist addition

* Added instructions for locally downloading YAML (#167)

* prepare for main

* prepare for merge to main

* add pre-puller for pipeline images

* Added instructions for downloading

* use dev branch instead of main

---------

Co-authored-by: Guillaume Moutier <[email protected]>
Co-authored-by: Guillaume Moutier <[email protected]>

* 03/06 and 05/05 pipelines working (#172)

* 03/06 and 05/05 pipelines working

* New image with openai deps

* Feature/streamline section 4 (#174)

* micor changes for clarity

* marking a chunk of section 04 as optional

* spellcheck

* Feat/rename sanity pipeline (#177)

* Updated the file name

* Images and guide updated

* Instructions + Working GitOps (#180)

* Fixed navigation (#195)

* Update workbench image version to 2.0.1 (rh-aiservices-bu#204)

* updating prepulled images to match 2.8.1 (rh-aiservices-bu#205)

* * hide diy (rh-aiservices-bu#206)

* fix the RH1-related stuff.
* formatting and bullet points
* specify user again

* Fix/sanity (rh-aiservices-bu#208)

* removing references to sanity.
* adding confidence pipeline image to puller. speeds up pipelines.
* also copied the image into a repo with a new name in quay.io

* change repo reference

* Fix #11 -  Revert model comparison to Flan-T5 (#26)

* Revert to Flan-T5

Remove CUDA from container

update packages and switch to CPU torch

update workbenches IS

update deployment

switch back to 1.3

update wb image

* update conclusion

* quick readme update (#27)

* adding parasol insurance throughout and updating titles (#28)

* Fix/remove auto (#30)

* removing auto from code

* adjusting instructions (pre-created is better than auto-created)

* git clone the right project

* adjust paths to new project name (#32)

* 05-05 changes branch to dev

* Added a step for the user to go to admin view (#37)

* fix service url of flanT5 (#39)

* WIP: Fixes (#40)

* what type of pipeline

* openAI clarifications

* let's not force it. users should see choice between kserve and modelmesh

* refreshed screenshots

* added jupyter notebook guide and rearrange order (#42)

* bump minio buckets to 101 (#44)

* added run button (#49)

* repo name in image fix

* Change userX to variable

* add token to vllm deployment (#53)

fix namespace

* New App (#54)

New look + RAG
Workench update
Content for part 3 - RAG

* add demo app (#55)

fix bootstrap

* adapt files for main

---------

Co-authored-by: Erwan Granger <[email protected]>
Co-authored-by: RHRolun <[email protected]>
Co-authored-by: adrezni <[email protected]>
Co-authored-by: rcarrata <[email protected]>
Co-authored-by: Cedric Clyburn <[email protected]>
Co-authored-by: RHRolun <[email protected]>

.github/.wordlist.txt

@@ -49,6 +49,8 @@ DevOps
 dex
 diag
 dir
+diy
+DIY
 dotenv
 drawio
 ds
@@ -107,6 +109,7 @@ Inlining
 insecureEdgeTerminationPolicy
 io
 ipynb
+jpg
 js
 json
 jsonpath
@@ -128,6 +131,7 @@ mAP
 md
 minio
 Minio
+MinIO
 mistralai
 mn
 mockups
@@ -141,6 +145,7 @@ nav
 Nodejs
 npm
 oc
+ocp
 ocs
 ods
 ojsonpath
@@ -179,6 +184,7 @@ pvc
 py
 pytorch
 Pytorch
+quantized
 rbac
 README
 readonly

README.md

@@ -1,10 +1,10 @@
-# Streamlining insurance claims with OpenShift AI
+# OpenShift AI Unleashed: Transforming Claims Processing for Maximum Efficiency!
 
 ## Introduction
 
-This repository contains the code, instructions, resources and materials associated with the Lab called **Streamlining insurance claims with OpenShift AI**.
+This repository contains the code, instructions, resources and materials associated with the Lab called **OpenShift AI Unleashed: Transforming Claims Processing for Maximum Efficiency!**.
 
-To consult the static version of the instructions, please use [this URL](https://rh-aiservices-bu.github.io/insurance-claim-processing/)
+To consult the static version of the instructions, please use [this URL](https://rh-aiservices-bu.github.io/parasol-insurance/)
 
 If you want to participate in the creation and update of this content, please consult the sections below.
 
@@ -62,7 +62,7 @@ If you want to install packages manually:
 
 ### Development
 
-From the main folder, launch `npm run dev`. This will launch both backend and frontend.
+From the main folder, launch `npm run dev` or `./start-dev.sh`. This will launch both backend and frontend.
 
 - Frontend is accessible at `http://localhost:9000`
 - Backend is accessible at `http://localhost:5000`, with Swagger API doc at `http://localhost:5000/docs`
@@ -114,9 +114,9 @@ done
 </details>
 
 <details>
-  <summary>Links for RH1 event environment assignment</summary>
+  <summary>Links for Summit event environment assignment</summary>
 
 - URL for all labs: [https://one.demo.redhat.com/](https://one.demo.redhat.com/)
-- Search for `insurance`
+- Search for `parasol`
 
 </details>

app/Containerfile

@@ -4,13 +4,13 @@ USER root
 
 WORKDIR /tmp
 
-RUN git clone https://github.com/rh-aiservices-bu/insurance-claim-processing.git
+RUN git clone https://github.com/rh-aiservices-bu/parasol-insurance.git
 
 FROM registry.access.redhat.com/ubi9/nodejs-18
 
 USER root
 
-COPY --from=stage /tmp/insurance-claim-processing/app/ /tmp/src
+COPY --from=stage /tmp/parasol-insurance/app/ /tmp/src
 
 RUN chown -R 1001:0 /tmp/src
 

app/Pipfile

@@ -0,0 +1,11 @@
+[[source]]
+url = "https://pypi.org/simple"
+verify_ssl = true
+name = "pypi"
+
+[packages]
+
+[dev-packages]
+
+[requires]
+python_version = "3.9"

app/README.md

@@ -0,0 +1,7 @@
+# Application image
+
+## Building
+
+The npm build happens during the image build. To do it successfully, you may have to augment the limits on open files in your system. Ex:
+
+`podman build --no-cache --ulimit nofile=10000:10000 -t rhoai-lab-insurance-claim-app:2.1.0 .`

app/backend/Pipfile

@@ -3,17 +3,30 @@ url = "https://pypi.org/simple"
 verify_ssl = true
 name = "pypi"
 
+[[source]]
+url = "https://download.pytorch.org/whl/cpu"
+verify_ssl = false
+name = "pytorch"
+
 [packages]
-boto3 = "~=1.28.0"
-fastapi = "~=0.104.0"
-httpx = "~=0.25.2"
-python-dotenv = "~=1.0.0"
+langchain = "==0.1.15"
+openai = "==1.17.0"
+pymilvus = "~=2.4.0"
+einops = "~=0.7.0"
+sentence_transformers = "~=2.6.1"
+boto3 = "~=1.34.81"
+fastapi = "~=0.110.1"
+httpx = "~=0.27.0"
+python-dotenv = "~=1.0.1"
 psycopg2-binary = "~=2.9.9"
-uvicorn = "~=0.23.2"
-python-multipart = "*"
+uvicorn = "~=0.29.0"
+websockets = "~=12.0"
+python-multipart = "0.0.9"
+torch = {version = "==2.2.2+cpu", index = "pytorch"}
 
 [dev-packages]
-"watchdog[watchmedo]" = "~=3.0.0"
+"watchdog[watchmedo]" = "~=4.0.0"
+black = "*"
 
 [requires]
 python_version = "3.11"