First version of node client

[m-chadda]

• pdf to image
• extension error
• cropping error
• getting iamges back, not working
• .
• pyscripts
• fixed scaling factor for page
• new compose
• scaling factor issues
• fixed pdla deploy
• added image url stuff :
• added image url stuff
• task creation has image_url
• rm threadlock and run in threadpool in pdla
• image file path fixes
• s3 upload issues
• done with ocr
• get tasks error fixed
• process lock on / route
• Refactored frontend
• Rebased with main succesfully
• Added OCR strategy and Task stats in viewer & status view
• main test
• removed segment density
• Moved dashboard to pages
• Added curl command placeholder for no tasks
• deployed new pdla
• added timing for ocr
• added timing for ocr
• table html to mkd
• table html to mkd
• table html to mkd
• Cleaned up svg errors, added oCR slection on upload form, images optimized
• fixing mkd table issue
• fixing mkd table issue
• adding \n as string literal
• adding \n as string literal
• | is not |?
• doing it on frontend
• doing it on frontend
• base
• frontend markdown resolved
• Rebasing with main
• Added tables markdown css, html viewing options, API key fixes for small screen menu dropdown
• updated google timeout
• list item fixes
• list item ol start from give digit
• clean ul
• migrations
• new migs
• new api
• new api spec
• wiping pg
• updating gpu
• updated gpu
• separated out task workers
• new workers - fast - highquality
• cleaned up pyscripts
• refactoring the models
• syntax error fix
• fixed syntax error
• fixed syntax error
• fixed pyscripts
• new dockers for processors
• new processor names
• cleanup
• Need to add chunking strategy in upload form
• Updating with main
• Added HTML viewing fixes and chunking strategy button on upload form
• No tasks state text prompt added
• task service docker created
• docker updated
• docker updated
• docker updated
• added uv venv
• added uv venv
• new kube
• paddle ocr issue
• created init
• paddle ocr issues, going to build with gpu
• added libgomp1 to dockerfile
• added openCV deps in docker
• updated num of workers
• repushing
• new kube
• unexpected end of data error
• dockerfile updated becuase cudnn error
• new kube
• updated env for docker image
• dockerignore update
• docker image inly copying over essentials
• docker image inly copying over essentials
• fixed COPY
• kube
• deploying new frontend
• task ocr parsed to float
• get content updated
• prints
• idek
• ready for deployment
• deploying new version
• deploying new version
• deploymetn with ocr
• deployment successful
• new cmds
• Fixed bug for API pricing calculator - fast pricing is now correct
• updated api spec
• Fixed billing amount bugs
• better tests
• stripe tested on dev
• invoicing done
• added authorization
• clean up
• rm unused code
• deployed
• token incorrect
• fixed incorrect server url
• rm cron from self deploy
• deployed newest version
• updated pdla
• updated pdla
• added
• added threadpool
• pdla code updated
• clean up services
• pdla fast new deployment
• deploying higher fast throughput
• updated number of connections
• invoicing fixes
• comment purge
• deployed invoicer update
• updated max dims for segments
• updated task
• cloudflare full strict
• Added small fixes for frontend and OCRBBox view on hover + OCRText for OCRBBox on hover
• Small fix - Removed random console.log
• updating queue consumer
• udpaed secrets
• updated deployment file
• Fixed MZ and Safari bugs for upload componenty on landing
• silly
• Removed speed approximations for Segmentation model cards
• file support
• deployed web
• silly
• added pollyfill for promise with resolvers
• libre office integration
• Fixed chunk_length undefined bug
• open api spec
• new venv
• Removed speed approximations for Segmentation model cards
• deployed web
• added pollyfill for promise with resolvers
• Fixed chunk_length undefined bug
• open api spec
• new venv
• sending to task service
• invalid header
• idk headers in pdfs suck
• updates to generate
• qwen
• qwen batch
• qwen batch
• qwen batch
• qwen batch
• qwen fixes
• qwen fixes
• qwen fixes
• qwen batch
• docx
• ppt, docx support added
• lfg
• lfg
• added task
• updated accept list
• merged
• fixed conflicts in qwen
• LLM based ocr for tables added
• updated html logic
• updated cost values
• fixed conflict in process
• fixed circular deps
• fixed circular deps
• updated default value
• removed print statements
• deploying new version with llm table support
• deployed new version with docx, pptx support
• added dropzone type
• added module declaration for react-dropzone
• deployed with LLM table support
• updated num_workers
• updated example kube env
• adding o1 support
• added latex api
• testing o1
• removed o1 stuff
• fixed monthly order for usage and multi file type support
• notices
• Update README.md
• Update LICENSE
• Delete LICENSE
• Create LICENSE
• Create COMMERCIAL_LICENSE
• bug fixes
• prompting
• Fixed free tier rollover on dashboard
• Api dialog width fix
• updated conversion
• Added file support byline
• added ppocr to table with llm
• camelot for tables
• deploying new web and backend
• adding o1 support
• added latex api
• testing o1
• removed o1 stuff
• fixed monthly order for usage and multi file type support
• bug fixes
• prompting
• updated conversion
• added ppocr to table with llm
• camelot for tables
• Small fix in byline for homepage
• adding sliding window to ppocr
• img to table
• img to table
• ocr result fix
• ocr result fix
• fixed pyscripts scaling issues and time logging
• fixed
• Added a lot of phone fixes
• moved file conversion to process, added pdf url to output
• fixed pyscripts and gen signed url
• rm prints
• Removed authentication from phone
• Fixed up pricing page header and styling
• cleanup - unused code
• cleanup - unused code
• default payment method
• new deployment
• cooked
• desc month
• month desc
• reverting
• reverting
• updated example secrets
• pdf_url
• added new logs for debugging
• removed logs
• adding logs to upload task
• newset version deployed
• changed api schema
• implemented new segment models in task
• update cropping
• Implmented new bbox stuff on frontend
• updated rust in docker
• deploying api schema change
• deploying api schema change
• added message for usage limit exceeded
• Using legacy version of pdfjs
• Added posthog analytics
• updated web
• Update README.md
• multi gpu support
• fixed typo for curl command
• fixed typo for curl command
• Update README.md
• Update README.md
• updated readme
• updated readme
• updated readme
• updated readme
• increased chunkr timeout
• increased timeout to 180s
• Updates
• added lazy static to client
• updated terraform
• add timeout for rrq
• server on fire
• server on fire
• lfg
• updating nodes to stop ooming
• seperated ocr from segmentation
• removed ocr lock
• dep hell
• throughput maxxing
• load model for every request with thread locks
• adding telemetry
• ocr more stable
• creating engine for each segment
• fixed syntax errro
• fixed syntax errro
• fixed syntax errro
• bentoml -> fastapi
• added pacakges to pyproject
• added logging
• accepting segments as str of json
• more logging
• more logging
• json 422 errors
• json 422 errors
• json 422 errors
• json 422 errors
• cleanup files
• ocr bangs
• pdf_file errors
• removed async from conversion
• cleanup
• issues with file conversion
• fastapi file response
• tempfile errors
• fixed for no conversion required
• to_pdf updated
• str -> path
• pdf path issues
• pdf path issues
• removed positional argument
• idk man
• background task for deletion
• ocr still fucky
• added queue for ocr engines
• ocr engine pooling
• logging
• Added copy button for copy/pasting markdown, html and JSON exactly as is. Can copy text from within chunks now as well.
• Got rid of extra console logs
• created ocr microservice
• using ocr service
• updated pyproject
• updated pyproject
• updated pyproject
• updated pyproject
• ocr test cmds added
• adding parrallel support for ocr
• adding parrallel support for ocr
• adding telemetry
• adding better telemetry
• adding better telemetry
• adding better telemetry
• adding better telemetry
• removed logs from ppocr
• updated dockers
• updated kube and docker
• removed torch
• updated dockers
• updated docker
• updated docker
• updated docker
• removed unused
• deps suck
• deployed new task and ocr services
• udpated timeouts
• gmi
• turning ocr pff
• OCR Off
• if no table html show table image
• segment chunk UI
• deployed web
• pdf file url implemented on web
• deployed web
• fixed bug with non pdf file viewing
• deployed web
• deployed web
• made /process async
• deployed
• ocr patch
• added multi OCR strategy
• idk
• added multi OCR config
• added multi OCR config
• added multi OCR config
• new login strat
• deploying
• new login
• new port
• new port
• new port
• new port
• new port
• new port
• new port
• deploying with aws login
• new port
• new port
• back to sq1
• cool
• changes to config
• changes to config
• changes to config
• changes to config
• changes to config
• adding aws cli
• deploying with aws textract support
• adding curl
• adding unzip
• update
• Update README.md
• task deployed
• installing into docker
• debugging
• added async context manager
• updating login
• Update README.md
• removed print statements
• Update README.md
• deployes task
• Update README.md
• Update README.md
• Update README.md
• Update README.md
• Added temp-banner for status updates to home page
• Update README.md
• deployed web
• deployed web
• removed out of date docker compose file
• Fixed textract bboxes on frontend
• Fixed OCR BBox stuff in python and for frontend
• deployed task
• rapid OCR slaps
• pushing to l4
• rapid ocr port change
• added cuda support
• added cuda support
• gpu support for rapid
• gpu support for rapid
• opencv
• adding unitable
• Fixed import in data.py
• Fixed imports trainer/utils.py
• Trying diff image
• Trying new image
• Removing text recognition
• It worksgit add -A
• added lock
• many workers
• many workers
• fixed ports on load balancer
• fixed ports on load balancer
• temp file cleanup
• garbage collection
• prints
• rm comments and prints
• adding unitable
• Fixed import in data.py
• Fixed imports trainer/utils.py
• Trying diff image
• Trying new image
• Removing text recognition
• It worksgit add -A
• adding uv to unitable
• update routes
• updated uv deps
• refactor
• rm modular ports
• gc collect in server
• added rapid ocr to kube
• updating rapidocr
• updating rapidocr
• added logging
• rm modular ports
• update readme
• rapid ocr deployed succesfully
• deployed rapidocr
• created server
• deployed task
• deployed task
• reverted task
• server created
• added python-dotenv
• throw error if no models specfied
• added logs
• added logs
• added logs
• getting first file
• downloading img as tempfile
• Image from bytes
• Image from bytes using bytesIO
• Image from bytes using bytesIO
• fixed syntax error
• spliting routes
• added html
• added html
• logging
• logging
• getting direct decode result
• unable to get html
• removed iPYTHON
• test
• better telemetry
• more logs
• more logs
• using tempfile
• loading model during inference
• html fix
• html fix
• improving html output
• improving html output
• adding logger
• adding logger
• adding logger
• error in response
• debugging bbox
• distributed queues
• handing off
• issues with file conversion
• rb
• updating pnpm lock:
• rebasing
• rebasing
• textractor is poorly written
• split into segmentation and ocr
• rebasing
• added paddle gpu
• fixed logger bug
• returning result
• returning result
• debugging
• debugging
• conversion errors
• conversion errors
• debugging
• response tpye json
• adding ororder
• ocr testing script works - with preprocessing
• ocr testing script works - with preprocessing
• mapping ppocr to html
• adding json output to server
• done - outputs are mid
• splitting
• done with unitable
• Preprocess worker done
• segmentation worker udpated
• preprocessing worker done
• added cropping
• added chunking
• segmentation done
• working on ocr worker
• preprocessing tested
• segmentation done
• rapid ocr output payloads created
• ocr done with rapid ocr
• rebasing
• rebasing
• debugging rapid ocr
• ocr works
• distributed postprocessing and segmentation
• queue splitting done
• connected table struct to rust
• done
• rename table ocr
• bbox output
• binarization
• constructing table
• preprocess
• preprocess
• merged
• merged
• confidence scores
• colspan row span
• table ocr implemented - some issues with response
• resize new table output
• resize new table output
• scaling issues
• updated services
• Preprocessing for ocr
• removed preprocessing of image
• Update README.md
• Update README.md
• Update README.md
• Updates
• updated dockers
• added table-ocr docker
• bruh
• removed model downloading from docker
• created kube .yaml files
• added pdfium binary to docker
• added wget to dockerfile
• building dockers
• preprocess copy error
• preprocess copy error
• fixed temp dir err
• preprocess copy error
• added col span and row span to cell
• semaphore on rapid OCR server
• added semaphore:
• updates
• updates
• updated deployments
• binary not found
• fix link
• fix link
• alright
• dokcer updated
• binary validation check
• fixed validation error
• new deployment
• better logs
• new deployment
• added file package to docker
• added file package to docker
• download with given file
• new deploy
• updated docker image
• new deploy
• 64but version of pdfium
• adding grafana
• new deploy
• adding prometheus
• not working out
• added monitoring values.yaml
• updated isDefaultDatasource
• added service monitor
• added rrq analytics
• reverting rrq
• added rrq api key
• new deployments for rrq analytics
• new deployments for rrq analytics
• new deployments for rrq analytics
• new deployments for rrq analytics
• ingress update rrq analytics
• deployment update rrq analytics
• deployment update rrq analytics
• new rrq
• deployed
• done
• table preprocessing
• ocr batching added
• udpated docker and new ocr
• logs on table ocr
• new deployments
• new deployments
• redistribute pods
• upgrade rrq
• gpu consumption
• cleanup
• cleaned up task service artifacts
• preprocess
• minor refactor and cleanup
• new rapid ocr models
• postprocessing done
• new models rapid ocr
• better models
• new rapid ocr models
• new models rapid ocr
• better models
• done
• gpu accelerator type
• cooked
• new deploys
• rapid ocr cooked
• throughput
• fixed engines
• engines suck
• rapid ocr 1 engine
• rm csvs
• new rapid ocr deployment
• updated rapid ocr
• new rapid ocr deployment
• fixed seralization error
• new rapid ocr deployment
• new deployments
• new analytics
• markdown tables fixed
• mkd down fix deployed
• testing changes
• Service down header removed
• Update README.md
• Update README.md
• Added github star and fork count to header
• Removed console.log
• new deploy
• new deploy
• pre apply discounts
• deployed
• Cahnges to github redirect button
• better logs
• new deploy
• new deploy: libreoffice fix
• deployment: github icon
• Small fix for header spacing
• adding github repo info
• removed pritn statements
• frontend updated
• deployed: github info with token
• page limits added
• fixed trigger
• deploy: page limits & discount bug fixes
• edits to usage
• Changed hardcoded limit values for fast and high quality on dashboard
• deploy
• deploy
• updated example secret
• updated self deployment
• updates
• adding paddle
• sends request
• updates
• bruh
• table ocr v1 done
• new table ocr
• adding compose.yaml
• cleaning up local dev
• compose yaml updated:
• issues with task expiration fixed
• postgres errors
• working on postgres migration errors
• works
• server working
• updated env
• updated env
• auto downloading pdfium binary works
• Update README.md
• wokring on pdla errors
• segemetation working locally
• wokring on table ocr
• testing options
• adding llm ocr
• working on LLM models
• added structured extraction
• uncooking merge errors
• rm unused code
• server - structured extract done
• done struc extract
• added kube for embeddings
• docker and kube for struc extract
• added structured extract to ocr off
• docker and kube for struc extract
• docker and kube for struc extract
• docker and kube for struc extract
• docker and kube for struc extract
• docker and kube for struc extract
• new dev deploy
• added structured extract to ocr off
• new kube for tei
• new ports
• done struc
• Frontend fixes for new output object
• cleaned up struc extract code
• docker and kube for struc extract
• updaing pyscripts
• updated pyscripts
• model updated
• hueristic for hallucinations
• cleanup
• cleanup
• new deployment
• merged properly
• updated pyscripts
• done with testing
• added paddle service
• added paddle service
• done with paddle docker
• docker compose updated
• deploy: web
• new prompts
• quality fixes structured extract
• new deployment
• updated pyscripts
• deploy: web
• table recognition working
• updated compose.yaml
• ocr done
• html to mkd v1
• html to mkd v1
• colspan works
• mkd done
• deploying new ocr
• deploying changes
• adding external s3 client to work with docker compose
• adding external s3 client to work with docker compose
• docker compose updated
• updated envs
• specfic dockers for different ocrs
• specfic dockers for different ocrs
• updated
• updated
• paddlex pipeline not working for table
• paddle proxy working
• added proxy to ocr services
• fixed glibc version error
• updated
• updated kube with proxy
• updated kube probe
• adding error handling in html to mkd
• deploy html fix
• ocr worker html safety
• fixing local dev
• new prompts
• quality fixes structured extract
• compose with vllm added
• updated
• kube experimental added
• readme updated
• updated readme
• updated readme
• updated readme
• updated readme
• updated readme
• Update README.md
• :wq Wq added stuff to prompt fot table ocr
• updated readme
• added .env.example
• updated LLM key in example.env
• added readme
• docker compose
• rm segmentation
• structured extract local works
• structured extract local works
• rm test
• new model
• very fast structured extract
• changes to configs
• configs updated
• added vllm support for tables
• updated table ocr
• updated prefix
• updated compose
• VLLM ocr works
• VLLM ocr works
• refactor
• table ocr with validation bangs
• works
• including prompts in binary
• formula prompt added
• tested
• added throttle limits
• cleaning up openai response
• cherrypicked
• experimental kube updated
• removed host rules
• bruh
• idk
• merged new seg strat
• merged table ocr with structured extraction
• built dockers and updated kube for tag d8eba61: cmd,fast,ocr,high,gen-ocr,post,pre,struc,table-ocr
• done
• done
• search config
• new kube
• filter out empty chunks
• new kube
• prompt and default value changes
• prompt and default value changes
• new cube
• updated openai call
• working on proxy
• using redis for rate limits
• compose changes
• leaky bucket rate limiter added with redis
• minor bug fixes
• minor bug fixes
• minor bug fixes
• paddle ocr test added
• rate limits implemented
• bugs
• deployment: rate limits
• deployment: rate limits
• kube
• bug fix
• table test
• added doctr service
• doctr inference server created
• adding doctr as ocr option
• doctr ocr done
• create new dockers
• docker not working
• fixed doctr docker
• deploy: doctr ocr
• deploy: doctr ocr
• new dockers
• doctr update
• doctr update
• doctr update
• doctr cant get nvidia drivers
• bulding docker with gpu
• new dockerimage for doctr
• new env example
• new secrets
• removed rrq message
• new sha
• added redis url
• rebasing
• testing more tables
• ocr integrated
• tested locally
• new kube for llm table ocr
• updated readme
• updated readme
• updated readme
• new kube for llm table ocr
• new kube for llm table ocr
• timeout issue
• ocr html validation removed
• new kube
• table ocr now fails on ocr error
• new docker build
• rrq timeout increase
• new dockers
• rebuild dockers
• Fixed formula rendering for html and markdown in viewer
• Final fixes
• presigned url headers
• compose
• new dockers and kube for pdf signed url
• removed dompurify
• deploying formula support on web
• Update compose.yaml
• page llm ocr
• tweaks to page prompt
• tweaks to page prompt
• view window
• tweaks to page prompt
• kube for page llm
• updated session timeout
• docs(README): fix incorrect git URL
• updated docker compose to use paddle ocr
• added comments
• page llm ocr
• tweaks to page prompt
• tweaks to page prompt
• view window
• tweaks to page prompt
• kube for page llm
• updated session timeout
• docs(README): fix incorrect git URL
• Update compose.yaml
• updated compose.yaml
• added jobs
• segmentation strategy:
• new redoc
• checkpointing tokenizer broken
• Sending workstation stuff
• segmentation strategy:
• new redoc
• expiration job created
• added comments for descriptions
• small changes
• checkpoint on segmentation forward pass
• new dockers
• added check to only delete finished tasks
• reading order and most prob segments done
• Fixed table colspan rowspan rendering
• terraform for azure
• renamed kube files to share across providers
• deleted old gcp kube - deprecated
• not required for internal communications
• removed backend config as that is only needed for ingress
• adding support for cloudflare tunnel
• cloudflare dns for tunnel and dns records
• cloudflare removed dsn records
• removed cloudflare terraform
• adding helm support
• added helm support for all the services with templates
• set up secrets for helm and internal communications
• updated shared secrets
• added readmes
• debugging
• adding redis to kube
• removed redis from terraform for azure
• added gpu support for embeddings
• added ngnix setup instructions
• adding cloudflare tunnel support
• cloudflared tunnel works
• updated envs
• kubernetes on azure
• clean up
• updated container name
• recovered gcp-experimental
• web bug fixes
• increase max nodes on azure
• adding postgres support
• renamed
• added postgress support with init db
• added local filesystem support for s3
• fixed bugs
• postgres permission issues fixed
• added pvc to redis, postgres and s3
• s3proxy works with azure blob storage
• added time slicing for gpus
• time slicing works
• Fixed table colspan rowspan rendering
• terraform for azure
• renamed kube files to share across providers
• deleted old gcp kube - deprecated
• not required for internal communications
• removed backend config as that is only needed for ingress
• adding support for cloudflare tunnel
• cloudflare dns for tunnel and dns records
• cloudflare removed dsn records
• removed cloudflare terraform
• adding helm support
• added helm support for all the services with templates
• set up secrets for helm and internal communications
• updated shared secrets
• added readmes
• debugging
• adding redis to kube
• removed redis from terraform for azure
• added gpu support for embeddings
• added ngnix setup instructions
• adding cloudflare tunnel support
• cloudflared tunnel works
• updated envs
• kubernetes on azure
• clean up
• updated container name
• recovered gcp-experimental
• web bug fixes
• increase max nodes on azure
• adding postgres support
• renamed
• added postgress support with init db
• added local filesystem support for s3
• fixed bugs
• postgres permission issues fixed
• added pvc to redis, postgres and s3
• s3proxy works with azure blob storage
• added time slicing for gpus
• time slicing works
• added backend s3 config for terraform
• updated readme
• updated backend config
• autoscale bug fixed
• updated deployment config
• deployment changes
• add configMap to overwrite the envs with envFrom
• added s3 backend to gcp
• minio successfully added
• updated web
• Fixed table colspan rowspan rendering
• terraform for azure
• renamed kube files to share across providers
• deleted old gcp kube - deprecated
• not required for internal communications
• removed backend config as that is only needed for ingress
• adding support for cloudflare tunnel
• cloudflare dns for tunnel and dns records
• cloudflare removed dsn records
• removed cloudflare terraform
• adding helm support
• added helm support for all the services with templates
• set up secrets for helm and internal communications
• updated shared secrets
• added readmes
• debugging
• adding redis to kube
• removed redis from terraform for azure
• added gpu support for embeddings
• added ngnix setup instructions
• adding cloudflare tunnel support
• cloudflared tunnel works
• updated envs
• kubernetes on azure
• clean up
• updated container name
• recovered gcp-experimental
• web bug fixes
• increase max nodes on azure
• adding postgres support
• renamed
• added postgress support with init db
• added local filesystem support for s3
• fixed bugs
• postgres permission issues fixed
• added pvc to redis, postgres and s3
• s3proxy works with azure blob storage
• added time slicing for gpus
• time slicing works
• struc extract changes
• basic VGT splitting
• embeddings on a100 by default
• updated time sliciing
• deploy: task expiration
• new dpeloyment
• removed api key
• refactor
• refactor
• server for vgt done
• rm pdla
• cleanups
• finalized vgt
• cooked
• cooked
• created server for reading order
• reading order server works
• outputs cooked
• reading order working with pairs
• added visualizer
• memory management for VGT
• experimenting with word grid
• mem management and reading order
• oom protection
• oom protection
• done reading order
• rm pr branch
• merged
• refactoring new worker with pipeline config
• new models created
• refactoring done with new api schema
• added first pipeline
• added first pipeline
• first 2 steps work
• clean up
• clean up
• clean up
• added mimetype to postgres
• ready for page processing
• ocr implemented
• rename
• deleted unused services
• deleted unused docker
• ocr auto and all done
• kill server management
• kill server management
• making segmentation models
• added merging and chunking
• rename step
• rename step
• updated chunking
• refactored in and out models for vgt server
• refactored in and out models for vgt server
• retries on ocr with backoff
• working on segmentation
• edit readme
• fix models
• refactor
• refactored clients to 1 file
• json.load array deserialization error
• segmentation works
• image processing added to upload form
• made high res images optional
• cropping works
• cropping done
• generating html in segment processing
• working on segment generation
• added macro for llm templates
• fixed typo
• fixed typo
• fixed typo
• segment processing done
• init for package
• added sci kit learn clustering
• cors
• cors
• working on upload model
• working on upload model
• added sci kit learn clustering
• added sci kit learn clustering
• added sci kit learn clustering
• added sci kit learn clustering
• added sci kit learn clustering
• added sci kit learn clustering
• reading order
• ocr and segmentation concurrently
• cropping issues
• added sci kit learn clustering v1
• segment processing works
• cors
• refactored rate limits
• Picture config added
• rewrite done
• added changelog
• fixed typo
• final reading order
• final reading order
• done
• expose threshold vgt
• upadted readme
• fixed dockerfiles
• refactor
• ok it works
• increase task replicas
• vgt
• adding cpu and mac support
• added cpu support
• new vgt +reading order again
• new vgt +reading order again
• updated gcp deployment:
• fixed cooked readme
• gcp deployment successful
• chunkr v1.1.0 services added
• fixed docker issue
• added cpu support
• added cpu support
• removed warning
• updated segmentation service
• connected improved segmentation with ocr
• updated dockers
• updated dockers
• updated changelog
• fix reading order edge case
• added libreoffice
• pdf plumber in tests
• single col pdf fix
• segmentation bug fix
• updated readme
• updated python client and added tests
• updated python client and added tests
• updated python client
• added better docstrings and improved imports
• removed unused imports
• improved docstrings
• removed ruff cache
• removed test.tsx
• added url and base64 support
• refactored
• deleted outdated code
• working on configuration testing
• ocr_strategy and expiration work
• chunkr client works
• json schema works
• readme updated
• publish new version of python client
• removed uv.lock
• renamed services
• removed unused model
• fixed random bugs
• fixed ocr mapping bug
• improved chunking
• new client published
• fix ol bug
• deployed bug fixes
• server
• return batch in doctr as completed
• Landing page improvements
• Working on features page
• Pricing container remaining on landing
• Bottom features and pricing section progress
• Bottom features and pricing section progress
• Bottom features and pricing section progress
• Starting dashboard overhaul
• Working on new dashboard
• Updates
• Skeleton completed for landing
• Changing features section
• Image deletion
• Remove tracked images
• Fixing commit history
• Fixing commit history
• Working on feature section
• Features section crads
• Moving features sections
• Pricing section sizing fixes
• Updates to pricing section and home layout
• Changed pricing cards
• Footer rework
• COmpleted landing page layout, styling and general skeleton. Cleanup, copy and finishing touches remaining
• Added MomentumScroll for homepage
• Dynamic additions
• Adding table for dashboard
• Dynamic table added
• Viewer adjustments for overhaul
• Working on latez rendering html view
• Completed viewer v1
• Viewer with detached panel infinite scroll
• v2 of viewer completed
• Fixed ref and horizontal press to scroll
• Toast functionality and segment specific json viewing and copying
• Small fixes for spacing sizing in Viewer
• Resizing handle bar style fixes
• Optimized loading times for viewer * table
• Optimized routing
• URL params optimized
• Working on Upload dialog
• Reverted formula messups and adding new upload component
• Small fixes
• Rebased succesfully with main (I think) - added some more nav
• Small sizing fix for file name tag
• Completed new upload component
• Added logic for usage fetching and processing
• Added taskbar vieing and better nav x
• New api management component started
• Completed viewer data piping
• Checkpoint - viewer completed
• Checkpoint - deleted older pages and ocmponents
• Checkpoint: Fixes to upload component for different states on DLA toggle
• Checkpoint: Task table almost completed (CRUD remaining)
• Working on structured extraction viewer
• Checkpoint: Extraction viewer complete
• Task table completed with crud
• Small fix for structured extraction viewing
• Fixed timeout between scrolls on viewer
• Commented out stuff from web for merging to main
• Commented out stuff from web for merging to main
• Fixes for svg build
• Fixes for build errors
• delete task done
• cancelling task works
• working on update task
• added direct methods
• ok update works
• updated migrations
• updating routes
• fixed bug in cancellation
• deployed with deletion bug fixed
• new Task struct
• updated task payload
• reconstructs pipline from artifacts on update
• added failure cases to task completion
• task processing
• struc
• new upload implementation
• 1 billi prompts
• 1 billi prompts
• update works
• update works
• segment llm prompting works
• removed dompurify
• added file to server
• deployed crud
• patch chunking
• Fix for tables in markdown view - switching to new branch
• Fixes for markdown tables, task table expiry date
• Small fixes
• Fixes for markdown table, task table expiry date and added messages to task table for frontend
• updated client
• cooked
• deployments
• small change to vgt server
• updated client
• add timeout as optional
• rebased main
• rebase
• max batch size for segmentation
• deployed new segmentation
• inference
• ?
• cooked
• reverted to old ocr
• new
• moved inside async lock
• updated time slicing
• stable deploy
• merged
• merged conflicts fixed
• cool
• added reranker
• new inference
• cool
• added reranker
• rust batching instead of batch/async
• moved entierly over to batching instead of dynamic batching
• new COMPOSE
• new helm
• new helm
• batch inference done
• filter out empty strings
• filter out empty strings
• se done
• updated embeddings
• add taskresponseasync to python client exports
• client updates to url
• done
• refactor task responses
• merged
• rebased
• async client cooked
• removed aiohttp import
• updated
• added docstring
• deleted unused scripts
• 1/2 reading order bug fixed
• Simplify reading order algorithm by removing column boundary checks and using only width ratio to identify wide elements.
• Created v1 node client
• First version of node client completed with simple tests
• First version of node client completed with simple tests
• Small fix for imports - publishing
• Completed first stable version of node client
• Changed certain models to classes and added fucntionality tests
• Updated package.json for devDependencies
• Rebasing with main