Commit History
fix: Prevent summary from overwriting first line of transcription
9cee14f
chore: Address Credo messages
59fddbd
perf: Combine `map` & `join` into `map_join`
91b179b
docs: Add @moduledoc for MedicalTranscription.Utilities
7c89e37
feat: Display weighting in code title; mixed color for mixed weighting
b9d0f1b
fix: Always pass `weighting` as a list
3d65418
chore: Include atom for each weighting result
10eac7e
feat: Weight based on multiple feedbacks for the same code
682498c
feat: Include codes with feedback, even if not most similar
0bce68f
refactor: Split code weighting into separate functions
69180e6
fix: Sort code results after weighting
af555df
chore: Remove unused code for highlighting previous feedback responses
b20fcc2
feat: Highlight weighted feedback
65843fa
feat: Weight results of code vector search
ea91b21
fix: Ensure we only return one feedback for a given text + vector pair
969feed
chore: Add timestamps to code_feedbacks and code_vectors tables
745fdb9
chore: Expand message on submitting code feedback
9aeba4b
feat: Highlight previous feedback in tag results
aaa724e
feat: Retrieve cosine similarity from Postgres query
076da4c
refactor: Create shared `collect_errors` function for Ecto changesets
609a959
chore: Rename Transcriber to prepare for a transcription context
4553650
refactor: Move CodeSearcher and CodeVector into Coding context
a7feaeb
refactor: Create feedback context
b01bac2
fix: Add namespace to serving in TranscriptionFilter
48ded1e
chore: Move audio pipeline handling to a GenServer
4a375b0
refactor: Remove need for `is_summary` flag
60bcf8d
chore: Shorten live audio chunk duration
139ee5c
refactor: Rename MembraneFilter to TranscriptionFilter
3e57477
docs: Document RecordingPipeline and Recorder modules
c574dff
chore: Remove dbg() call in audio pipeline
3c0f1de
chore: Extract recording pipeline into separate module
3e013eb
chore: Move child specs to functions in AudioTagger
4b67cec
chore: Namespace serving child processes
a227b91
feat: Batch call to serving for text embeddings
69f8a70
chore: Enable Ecto repo
02f6fff
chore: Add text embedding serving to supervision tree
9c1c3b1
feat: Search for similar embeddings in DB
d7e3459
chore: Separate number with underscore and add moduledoc
95e1261
chore: Add release function to precompute code vectors
10ceb35
chore: Move vector precomputation into separate module
fd1b984
fix: Pass model_tuple for tagging full summary transcription
3c07b80
feat: Remove unnecessary GenServer for building code vectors
847217a
feat: Store vector for text in feedback, associate to code_vectors
58815a9
feat: Search for similar embeddings in DB
f23612f
feat: Create `code_vectors` table
b3d858d
docs: Add moduledocs to some modules
cb8dc08
chore: Configure for connecting to Neon's Postgres
f22b0a2
code review feedback
99149d9
Chris Held
commited on
tag full transcription text
96c379c
Chris Held
commited on