CourtListener PipeDocs captured as the INPUT to the chunker step — one gzipped ai.pipestream.data.v1.PipeDoc per source doc. Drop-in stress / regression fixture for the chunker module.
CourtListener PipeDocs captured as the INPUT to the embedder step — one gzipped ai.pipestream.data.v1.PipeDoc per source doc. Drop-in stress / regression fixture for the embedder module.