feat: Generate vector embeddings from UI via GenServer

Dockerfile

@@ -68,7 +68,7 @@ RUN mix release
 FROM ${RUNNER_IMAGE}
 
 RUN apt-get update -y && \
-  apt-get install -y libstdc++6 openssl libncurses5 locales ca-certificates \
+  apt-get install -y libstdc++6 openssl libncurses5 locales ca-certificates curl unzip \
   && apt-get clean && rm -f /var/lib/apt/lists/*_*
 
 # Set the locale

lib/medical_transcription/application.ex

@@ -29,7 +29,8 @@ defmodule MedicalTranscription.Application do
       # Start a worker by calling: MedicalTranscription.Worker.start_link(arg)
       # {MedicalTranscription.Worker, arg},
       # Start to serve requests, typically the last entry
-      MedicalTranscriptionWeb.Endpoint
+      MedicalTranscriptionWeb.Endpoint,
+      MedicalTranscription.Vectorizer
     ]
 
     # See https://hexdocs.pm/elixir/Supervisor.html

lib/medical_transcription/vectorizer.ex

@@ -0,0 +1,52 @@
+defmodule MedicalTranscription.Vectorizer do
+  use GenServer
+
+  # Client API
+  def start_link(_) do
+    GenServer.start_link(__MODULE__, :pending, name: MedicalTranscription.Vectorizer)
+  end
+
+  # Server callbacks
+  @impl GenServer
+  def init(state) do
+    IO.inspect(state, label: "Vectorizer GenServer started with state")
+
+    {:ok, state}
+  end
+
+  @impl GenServer
+  def handle_call(:get_status, _from, state) do
+    # Return an atom containing the current status
+    {:reply, state, state}
+  end
+
+  @impl GenServer
+  def handle_cast(:generate, state) do
+    if state == :pending do
+      send(self(), :generate)
+
+      {:noreply, :in_process}
+    else
+      # Do nothing
+      {:noreply, state}
+    end
+  end
+
+  @impl GenServer
+  def handle_info(:generate, _state) do
+    # Build the embeddings
+    IO.puts("Building vectors embeddings for ICD-9 codelist")
+
+    cached_download =
+      AudioTagger.SampleData.cache_dir()
+      |> Path.join("icd9_codelist.csv")
+
+    unless File.exists?(cached_download) do
+      AudioTagger.SampleData.get_icd9_code_list_csv()
+    end
+
+    AudioTagger.Vectors.precalculate(cached_download)
+
+    {:noreply, :complete}
+  end
+end

lib/medical_transcription_web/components/layouts/app.html.heex

@@ -9,10 +9,20 @@
       </h1>
     </div>
 
-    <div class="px-6">
+    <div class="flex-1 px-6">
       <p class="text-xs leading-normal tracking-[0.2em] font-semibold uppercase">Today</p>
       <!-- TODO: Show a history of uploaded files here. -->
     </div>
+
+    <div class="flex flex-col items-center">
+      <button
+        class="px-3 py-2 bg-brand text-white rounded-lg"
+        type="button"
+        phx-click="build_vector_embeddings"
+      >
+        Build vector embeddings
+      </button>
+    </div>
   </header>
 
   <main class="flex-1 pl-16 pr-16 pt-[25px]">

lib/medical_transcription_web/live/home_live/index.ex

@@ -110,6 +110,13 @@ defmodule MedicalTranscriptionWeb.HomeLive.Index do
     {:noreply, assign(socket, :status, :pending)}
   end
 
+  @impl Phoenix.LiveView
+  def handle_event("build_vector_embeddings", _params, socket) do
+    GenServer.cast(MedicalTranscription.Vectorizer, :generate)
+
+    {:noreply, socket}
+  end
+
   @impl true
   def handle_info({:transcription_row, chunk_result}, socket) do
     # The processing sends a message as each chunk of text is coded. See here for some background and potential

mix.lock

@@ -1,5 +1,5 @@
 %{
-  "audio_tagger": {:git, "https://github.com/headwayio/audio_tagger.git", "bf49e4d3cd86e0b7123423595818d23b1967ca69", []},
+  "audio_tagger": {:git, "https://github.com/headwayio/audio_tagger.git", "70f8ce8abe45b4f26ddd3876bc935845823c3f8c", []},
   "aws_signature": {:hex, :aws_signature, "0.3.1", "67f369094cbd55ffa2bbd8cc713ede14b195fcfb45c86665cd7c5ad010276148", [:rebar3], [], "hexpm", "50fc4dc1d1f7c2d0a8c63f455b3c66ecd74c1cf4c915c768a636f9227704a674"},
   "axon": {:hex, :axon, "0.6.0", "fd7560079581e4cedebaf0cd5f741d6ac3516d06f204ebaf1283b1093bf66ff6", [:mix], [{:kino, "~> 0.7", [hex: :kino, repo: "hexpm", optional: true]}, {:kino_vega_lite, "~> 0.1.7", [hex: :kino_vega_lite, repo: "hexpm", optional: true]}, {:nx, "~> 0.6.0", [hex: :nx, repo: "hexpm", optional: false]}, {:polaris, "~> 0.1", [hex: :polaris, repo: "hexpm", optional: false]}, {:table_rex, "~> 3.1.1", [hex: :table_rex, repo: "hexpm", optional: true]}], "hexpm", "204e7aeb50d231a30b25456adf17bfbaae33fe7c085e03793357ac3bf62fd853"},
   "bimap": {:hex, :bimap, "1.3.0", "3ea4832e58dc83a9b5b407c6731e7bae87458aa618e6d11d8e12114a17afa4b3", [:mix], [], "hexpm", "bf5a2b078528465aa705f405a5c638becd63e41d280ada41e0f77e6d255a10b4"},