ArchAIve

ArchAIve is an AI-powered artefact digitisation platform for the preservation and proliferation of heritage and culture. Designed and developed for the Singapore Chinese Chamber of Commerce and Industry (SCCCI) as part of the SCCCI x NYP collaboration in the NYP 25S1 IT3100 AI Applications Project module.

We are enthralled to bring to you ArchAIve, the fruition of months of our team's efforts. ArchAIve features specialised ML models, highly intelligent AI pipelines, nuanced business logic execution, and a smooth user experience. With ArchAIve, research teams can save a lot of time in digitising artefacts. ArchAIve is able to generate rich metadata information for freshly uploaded artefacts, featuring handwritten traditional Chinese calligraphy transcription, image detection, image captioning and face recognition. Safeguards, a human-in-the-loop approach and more efforts make the entire experience highly robust, safe and intuitive.

To find out more, check out the organisation README. This is a release package for the Backend repository, check out the README here.

Auto-generated PR Release Notes:

Added reqs and ignored models air by @Prakhar896 in #1
Added ModelStore, ThreadManager, FireStorage APIs by @Prakhar896 in #5
Updating personal branch with commits on main by @Prakhar896 in #7
New LLMInterface and model loading through callbacks by @Prakhar896 in #9
New LLMInterface and fixes for ArchSmith by @Prakhar896 in #12
Added Chinese Calligraphy Recognition (CCR) pipeline. by @JunHammy in #13
Added Binary Classifier (CNN) pipeline by @ZyuT0h in #14
ImageCaptioning pipeline and Preliminary ASVisual service by @Prakhar896 in #19
Completed ASVisualiser for better debugging by @Prakhar896 in #22
Added transcription processor class. by @JunHammy in #21
Added Named Entity Recognition (NER) Pipeline by @erpanter in #23
FileManager, ArchSmith Autopersistence and ASVisualiser Improvements by @Prakhar896 in #28
Added Image Enhancement (Real ESRGAN) class by @ZyuT0h in #27
Added MMProcessor, standardised error handlings by @JunHammy in #26
Added HF Processor by @ZyuT0h in #32
Added User model by @Prakhar896 in #36
Changed NER Pipeline to use pretrained model. Added CDN with image retrieval endpoint. by @JunHammy in #37
Identity workflow, Decorators and JSONRes Improvements by @Prakhar896 in #39
Added Artefact, Metadata, MMData, HFData Classes at models.py by @ZyuT0h in #38
Updated CORS Policy by @Prakhar896 in #46
Added Book and Batch Class by @ZyuT0h in #45
Updated Batch Model for Lower Computation by @Prakhar896 in #49
Category, CategoryArtefact models and 'id' attribute bug fix by @Prakhar896 in #53
Doc Strings for Data Models (Artefact, Metadata, HFData, MMData, Book) in models.py by @ZyuT0h in #57
DataImportProcessor, Threadsafe Persistence, Batch suite of models by @Prakhar896 in #58
Fixes for DataImportProcessor by @Prakhar896 in #61
Added In-item chatbot by @ZyuT0h in #60
Switched to NLTK NER, Improved CDN & Catalogue Retrieval, Added DB Utilities by @JunHammy in #63
Automatic Face Recognition, Updated HFProcessor, New Figure and Face models by @Prakhar896 in #65
New User-associated AuditLog model by @Prakhar896 in #68
Data Import API by @ZyuT0h in #66
User Profile Management, Upgraded enforceSchema, New EmailCentre by @Prakhar896 in #70
Reorganisation for DB schemas, API routes and Extensions utility by @Prakhar896 in #71
Added artType to Artefact model and artefactMetadata endpoint; enhanced getFaceImage API by @JunHammy in #72
Admin Blueprint, Profile Info in CDN, User load and save improvements by @Prakhar896 in #74
Added new API endpoints in dataImport.py by @ZyuT0h in #73
Modified confirmBatch to prevent starting a new batch by @ZyuT0h in #76
More IAM admin endpoints and email alerts by @Prakhar896 in #77
Added update artefact endpoints by @JunHammy in #80
Revised user data validation, logic fixes by @Prakhar896 in #82
Auto Categorisation workflow and CatalogueIntegrator service by @Prakhar896 in #84
Added LLM Inference for CCR Pipeline, enhanced endpoints with cache invalidation by @JunHammy in #86
LLM Inference Compatibility for Image Classification and Captioning by @Prakhar896 in #89
Seed Profiles Algorithm by @Prakhar896 in #92
Added cdn endpoints getCollectionMemberIDs and getCollectionDetails , enhanced getAllCategoriesWithArtefacts endpoint by @JunHammy in #91
Added getBatchFirstArtefact to cdn.py by @ZyuT0h in #95
Start integration now invokes HFCatalogueIntegrator by @Prakhar896 in #96
Figure Retrieval and Update Endpoints by @Prakhar896 in #98
Added chatbotBP for initemchat by @ZyuT0h in #99
Added getAssociationInfo and updateAssociation endpoints by @JunHammy in #102
Fixes for batch processing bugs, stage lifecycle downscaling by @Prakhar896 in #105
Added group management APIs by @JunHammy in #108
Provision of vetting status in metadata retrieval by @Prakhar896 in #112
Updated initemChat.py by @ZyuT0h in #111

Full Changelog: https://github.com/ArchAIve-Project/Backend/commits/v1.0

Note: Though Nicholas (@erpanter) has been listed as a contributor, he had left the team, and his work has thus been overwritten. Due to technical limitations with GitHub and commit history, he is falsely listed, but he is not a contributor, nor a part of The ArchAIve Team.