v0.4.4

What's Changed

• [completions and embeddings] Add support for Ollama (local LLMs) by @eolivelli in #711
• [examples] Finalise the Ollama example application by @eolivelli in #713
• Updated Azure doc ingestion example to include RAG chatbot by @devinbost in #712
• Add gRPC topic producer endpoint - Java side by @cbornet in #693
• Minor cleanups of AgentContext by @cbornet in #719
• Apps: implement force deletion and improve status by @nicoloboschi in #714
• [HuggingFace] Make hugging face embeddings work in docker run with multiple pipelines by @eolivelli in #721
• [docs] Improved readme by adding an important screenshot and description by @devinbost in #720
• Bump OpenAI to v1 by @cbornet in #723
• Use Python asyncio gRPC by @cbornet in #722
• Fix Pinecone config vector.metadata.X by @nicoloboschi in #726

Full Changelog: v0.4.3...v0.4.4

v0.4.3

What's Changed

• [cleanup] Put back Pulsar classes to the pulsar package by @eolivelli in #685
• CLI: option to customize the command name and version by @nicoloboschi in #686
• Upgrade Lombok to 1.18.30 to support building on JDK21 by @eolivelli in #687
• [langserve-invoke] Support non map response in /invoke by @eolivelli in #688
• Add examples and utility function to deal with Cassio by @eolivelli in #689
• [cli]] updates for multi-os browser for langstream UI by @mark878 in #682
• Complete python gRPC requests when closing by @cbornet in #690
• [cli] Add Commands to create an application from an archetype by @eolivelli in #691
• Change log level for Python environnment variables by @cdbartholomew in #694
• [python agents] Allow to access resources configured in configuration.yaml by @eolivelli in #692
• Metrics: add basic metrics about sources and sinks by @eolivelli in #696
• MiniLangstream should error out if no NodePort exists for a service (#698) by @Gagravarr in #700
• LangStream Home from Property, support Symlinks by @Gagravarr in #699
• [openai] Remove debug logging by @eolivelli in #702
• MiniLangStream enable NodePort if missing for services by @Gagravarr in #701
• [examples] Improve the S3 Source example by @eolivelli in #706

New Contributors

• @Gagravarr made their first contribution in #700

Full Changelog: v0.4.2...v0.4.3

v0.4.2

What's Changed

• [agents] Add support for exporting custom metrics to Prometheus by @eolivelli in #681
• [metrics] Add OpenAI metrics about chat completions and text completions by @eolivelli in #684
• Fix service gateway with topics by @nicoloboschi in #683

Full Changelog: v0.4.1...v0.4.2

v0.4.1

What's Changed

• [examples] correct misspelling in README.md by @mark878 in #677
• [streaming] Add support for Pravega.io as messaging broker by @eolivelli in #668
• [python] Add CLI command to run tests by @eolivelli in #680
• [langserve-invoke] Extract the LangServe Java Client code and Allow to parse simple strings and not only JSON by @eolivelli in #676
• [webcrawler] Handle bad sitemaps and allow to skip them by @eolivelli in #679

Full Changelog: v0.4.0...v0.4.1

v0.4.0

What's Changed

• Remove Pulsar ack timeout by @cbornet in #633
• Add e2e test for cli with docker run by @nicoloboschi in #631
• Automate brew update after release by @nicoloboschi in #632
• Add post-release test for brew and CLI by @nicoloboschi in #636
• Improve CLI E2E test by @nicoloboschi in #637
• Add possibility to change subscription type by conf by @cbornet in #634
• Add ability to provision persistent disks to agents by @eolivelli in #590
• WebCrawler: use disk to store status if configured by @nicoloboschi in #643
• Add example application about how to write to Slack using the http-request agent by @mark878 in #628
• CLI UI: start on port 8092 by default by @nicoloboschi in #647
• [examples] Add a chatbot built with LangChain and Astra DB by @eolivelli in #646
• [agents] Add support for using Camel components as sources by @eolivelli in #644
• Add sample integration with LangSmith by @eolivelli in #648
• Add persistent state directory to python agents by @nicoloboschi in #645
• Archetypes - part 1 by @eolivelli in #642
• [python] Add a CLI command to import pip requirements.txt by @eolivelli in #652
• [examples] Added new flow for azure-document-ingestion and terraform by @devinbost in #585
• [cassandra] Handle errors due to missing parameters by @eolivelli in #653
• [python] Add support for hot reload of python code in docker mode by @eolivelli in #654
• [ci] Unblock CI temporary by moving to OpenJDK instead of Temurin by @eolivelli in #659
• [webcrawler] Do not crash (restart the pod) in case of network glitches by @eolivelli in #657
• [docker] Do not close classloaders, to prevent spammy ClassNotFoundExceptions by @eolivelli in #656
• Gateway: accept HTTP request for producing messages by @nicoloboschi in #655
• [python] Do not crash in case of syntax errors in the files during DEV mode by @eolivelli in #658
• [python] use vfork to start the python interpreter (needed for openjdk) by @eolivelli in #661
• [tests] Use a fixed version of RedPanda that is known to work by @eolivelli in #662
• Update UIAppCmd.java to set UI on 0.0.0.0 by @mark878 in #667
• New gateway 'service' - part 1 by @nicoloboschi in #663
• [python] Add new agent type python-service by @eolivelli in #665
• API gateway: cache topic producer for the same gateway by @nicoloboschi in #660
• CLI: add .env configuration file by @nicoloboschi in #670
• [agents] Add Agent to call LangServe (langserve-invoke) by @eolivelli in #673
• Improved example and resolved issues with Terraform deployment script by @devinbost in #674
• Service gateway: direct interaction with agent by @nicoloboschi in #669
• Fix WebCrawler state-storage default value by @nicoloboschi in #675

New Contributors

• @mark878 made their first contribution in #628
• @devinbost made their first contribution in #585

Full Changelog: v0.3.1...v0.4.0

v0.3.1

What's Changed

• CLI: fix temp file permissions by @nicoloboschi in #629

Full Changelog: v0.3.0...v0.3.1

v0.3.0

What's Changed

• Handle message formatting to support markdown and code blocks by @mfortman11 in #567
• [vector-databases] Add support for Apache Solr by @eolivelli in #565
• Adding example of web crawling chatbot using RAG and chat memory by @cdbartholomew in #568
• Remove Pulsar dependencies from agents-commons and ai-agents by @cbornet in #569
• React-Chatbot-UI: Add Loading state and style updates by @mfortman11 in #576
• Upgrade pyhon deps by @cbornet in #555
• Rename SchemaSerializationException to SerializationException by @cbornet in #577
• Update variable format for chatbot-rag-memory example by @cdbartholomew in #573
• Add opensearch support by @nicoloboschi in #574
• Fix pulsar runtime depency conflicts by @cbornet in #580
• Fix gateway with Pulsar by @cbornet in #586
• [compute-ai-embeddings] Rework support for Hugging Face local models for embeddings by @eolivelli in #584
• Support gateway start from partitioned topics offsets by @cbornet in #588
• Apache Solr support: implement DELETE collection by @eolivelli in #589
• CLI: fix Mermaid generation without gateways by @nicoloboschi in #591
• Cherry-pick OpenSearch with patched nar plugin by @nicoloboschi in #593
• CLI UI: do not print stacktrace if 'open' command is not available by @nicoloboschi in #594
• Add vector database sink config validation and documentation by @nicoloboschi in #583
• Add severity filter for logs to UI by @cdbartholomew in #598
• Improve missing AI service validation by @nicoloboschi in #597
• CLI docker run: create a dedicated profile by @nicoloboschi in #596
• Docker run mode: fix Mermaid rendering and the default profile by @eolivelli in #601
• Add e2e tests for gateway with Pulsar by @cbornet in #602
• Fix runtime image size by @nicoloboschi in #603
• [query agent] Allow to execute DML statements by @eolivelli in #605
• [examples] Add a sample chat application with PostGRESQL supporting a chat history by @eolivelli in #607
• Initial support for AWS Bedrock LLM (embeddings and completions) by @nicoloboschi in #599
• Add support for topic schemas in Pulsar by @cbornet in #608
• Add more flow control agents: timer-source and trigger-event by @eolivelli in #609
• [docker run] Exit the container in case of errors by @eolivelli in #615
• Move OpenSearch configuration to use a single index by @nicoloboschi in #621
• Add Rag with AWS example by @nicoloboschi in #617
• Remove deadLetterTopicProducer from Kafka consumer conf by @cbornet in #620
• Fix Kafka topic options not set by @cbornet in #619
• [query agent] Add test case with output-field that contains a dash character by @eolivelli in #616
• Add support for DLT with Pulsar by @cbornet in #618
• CLI: cache github repository for faster deployments by @nicoloboschi in #622
• Cli: fix docker run on linux by @nicoloboschi in #604
• Add possibility to set retention policies in Pulsar by @cbornet in #624
• Add example instance.yaml for Pulsar by @cbornet in #626
• Add tests for Pulsar retention setting by @cbornet in #625
• Helper scripts to set the REACT_APP environment variables by @cdbartholomew in #614
• Enhance the example applications: chat history and webcrawler (and add the 'log-event' agent) by @eolivelli in #623
• Add suport for KeyValue in Pulsar by @cbornet in #627

Full Changelog: v0.2.0...v0.3.0

v0.2.0

What's Changed

• Fix OpenAI env var names in READMEs by @cbornet in #528
• Add an example on how to implement websocket clients (javascript) by @nicoloboschi in #527
• Cli: include .bat file in the tarball by @nicoloboschi in #530
• OpenAI: implement logprobs (preparation for the FLARE pattern) by @eolivelli in #531
• [agents] New agent 'dispatch' for message routing by @eolivelli in #536
• Add support for PDF, doc, ppt, and other file types in Python/LangChain by @cbornet in #539
• Add dry-run apps deployment by @nicoloboschi in #533
• Fix placeholder resolution for non string fields by @cbornet in #550
• Add example using a configurable LangChain document loader by @cbornet in #543
• New http-request agent by @nicoloboschi in #537
• Add Azure blob storage source by @nicoloboschi in #546
• Add e2e test for S3 source by @nicoloboschi in #552
• Fix WebCrawler handle-cookies option by @nicoloboschi in #553
• New UI for apps monitoring and testing by @nicoloboschi in #551
• Improve Langchain document example by @cbornet in #557
• Improve apps logs command for CrashLoopBack and add JSON format by @nicoloboschi in #559
• Apps UI: support chat gateways by @nicoloboschi in #560
• Upgrade jandex maven plugin by @nicoloboschi in #558
• Add Chatbot UI to examples by @mfortman11 in #556
• Implementations to build the Flare (Forward-Looking Active REtrieval augmented generation) pattern with LangStream by @eolivelli in #548
• Switch the env var prefix to REACT_APP by @cdbartholomew in #563
• CLI UI: add mermaid diagram by @nicoloboschi in #562
• App UI: escape HTML and render markdown (optional) by @nicoloboschi in #566

New Contributors

• @mfortman11 made their first contribution in #556

Full Changelog: v0.1.0...v0.2.0

v0.1.0

What's Changed

• Switch Python agents with new gRPC-based ones by @cbornet in #488
• Simplify readme by @nicoloboschi in #498
• [apps] Allow to specify default values for globals in configuration.yaml by @eolivelli in #499
• [breaking] Refactor Python user API by @cbornet in #495
• [docker-run] Allow to not pass secrets by @eolivelli in #501
• Add herddb to mini-langstream by @nicoloboschi in #503
• E2e Tests: allow to pass google json sa by @nicoloboschi in #496
• [ai-agents] Introduce the re-rank agent with MMR ranking by @eolivelli in #502
• Add chart to docker-chatbot example by @cbornet in #505
• Kafka: add serializer for raw Maps and Lists by @eolivelli in #506
• Introduce framework to automatically generate doc for agents by @nicoloboschi in #504
• Add doc for Gen AI toolkit agents by @nicoloboschi in #508
• Document Kafka Connect, text-processing, vector, re-rank agents by @nicoloboschi in #509
• [breaking change] Do not use Mustache but SpringBoot like references to refer to secrets and globals by @eolivelli in #507
• Document webcrawler source and python by @nicoloboschi in #511
• [compatibility] Backward compatibility with references to secrets and globals using the Mustache syntax by @eolivelli in #515
• Generate documentation for resources (AI services, datasource, vector-database) by @nicoloboschi in #513
• [webcrawler] Fix NPE in case of wrong URLs without slashes by @eolivelli in #518
• Generate doc for assets by @nicoloboschi in #517
• Move TransformContext to a dedicated module by @eolivelli in #519
• Update docker-chatbot chart by @cbornet in #521
• Handle compatibility with 0.0.x syntax for prompts and embeddings by @eolivelli in #520
• Improve e2e tests stability by @nicoloboschi in #525

Full Changelog: v0.0.22...v0.1.0

v0.0.22

What's Changed

• Add e2e test with WebCrawler + embeddings + vector sink + completions by @nicoloboschi in #444
• Add Python processor grpc server by @cbornet in #440
• [logging] Do not print secrets while opening a connection to the Vector database by @eolivelli in #446
• E2E tests: allow to use an external kafka cluster by @nicoloboschi in #450
• Add Java gRPC Source by @cbornet in #449
• CLI: Allow updates to the default profile by @nicoloboschi in #452
• [cli] Print a log message before starting to clone a git repo by @eolivelli in #451
• Add gRPC Source in Python server by @cbornet in #455
• E2E tests: allow to use an external code storage by @nicoloboschi in #456
• Do not display transfer progress during CI by @cbornet in #459
• Add integration test for Python Source by @cbornet in #458
• E2E tests: setup control plane authentication by @nicoloboschi in #462
• Add permanent failure to Python gRPC Source by @cbornet in #461
• Allow to return dict instead of Records in Python agents by @cbornet in #460
• Limit the number of read records to prevent OOME by @cbornet in #463
• Add some Cassandra tests with Vector type by @eolivelli in #465
• [kafka] Fix passing the replication-factor to create topics by @eolivelli in #464
• Add Java gRPC Sink by @cbornet in #466
• Update README.md by @cdbartholomew in #467
• [compute-ai-embeddings] Compute Embeddings Agent: execute calls asynchronously and in batches while preserving per-key ordering by @eolivelli in #457
• Mini langstream by @nicoloboschi in #468
• [webcrawler] Do not fail in case of unknown properties in the JSON file (handle backward compatibility) by @eolivelli in #469
• [vector-db-support] Add support for Milvius.io - part 1 by @eolivelli in #470
• Fix compile warning by @nicoloboschi in #474
• Executor pod: set liveness and readiness probe by @nicoloboschi in #475
• Improve mini-langstream packaging by @nicoloboschi in #473
• E2E tests: make app resources customizable by @nicoloboschi in #476
• [milvus] Add support for Milvus.io - part 2 by @eolivelli in #472
• Add gRPC Sink Python server side by @cbornet in #478
• Implement ai-text-completions - OpenAI by @nicoloboschi in #477
• [milvus] Implement create-index command by @eolivelli in #480
• Fix app update and add test by @nicoloboschi in #479
• Add support for Milvus Cloud - Zillis by @eolivelli in #485
• Implement ai-text-completions for VertexAI and HuggingFace inf by @nicoloboschi in #484
• Fix ai-text-completions example by @nicoloboschi in #483
• Add e2e test for Python Sink by @cbornet in #482
• [vector-db-sink] Implement support for writing to JDBC services by @eolivelli in #486
• Allow to configure more than one AI service by @nicoloboschi in #489
• [jdbc] Implement the "jdbc-table" asset and add "--start-database" to "docker run" mode by @eolivelli in #490
• Fix VertexAI parameters by @nicoloboschi in #491
• [local mode] Add examples about how to deploy HerdDB in minikube by @eolivelli in #493
• Load app code in the setup job by @nicoloboschi in #492

Full Changelog: v0.0.21...v0.0.22