Fixed - Fixed `completions` command that panicked after we upgraded to new `clap` version
0.93.0
Added - By default, we will resolve IPNS DNSLink URLs (e.g. `ipns://dataset.example.org`) using DNS query instead of delegating to the gateway. This is helpful when some gateway does not support IPNS (e.g. Infura) and in general should be a little faster and provides more information for possible debugging
0.92.0
Added - `kamu system ipfs add` command that adds dataset to IPFS and returns the CID - it can be used to skip slow and unreliable IPNS publishing Fixed - Engine runtime error when it was incorrectly using a `.crc` file instead of a parquet data file
0.91.0
Changed - Clean up cached data files upon successful commit to save disk space - Push `cache` directory of datasets to remote repositories - This is a temporary measure to allow resuming heavy-weight ingest actions from checkpoints
0.90.0
Added - Experimental support for templating of ingest URLs with environment variables: yaml fetch: kind: url url: "https://example.org/api/?apikey=${{ env.EXAMPLE_ORG_API_KEY }}"
- Experimental support for containerized fetch steps: yaml fetch: kind: container image: "ghcr.io/kamu-data/example:0.1.0" env: - name: SOME_API_KEY
Changed - Flag `kamu pull --force-uncacehable` was renamed to `--fetch-uncacheable`
0.89.0
Added - When pushing to IPNS execute `ipfs name publish` even when data is up-to-date to extend the lifetime of the record.