Show error when not enough points to calculate umap

index.html

@@ -41,7 +41,8 @@ <h2 class="mb-0" style="font-weight: 300">SIMS Web</h2>
                         <div class="custom-file">
                             <input type="file" class="custom-file-input" name="file_input" id="file_input"
                                 accept=".h5ad">
-                            <label id="file_input_label" class="custom-file-label" for="image">Select an AnnData/Scanpy
+                            <label id="file_input_label" class="custom-file-label" for="file_input">Select an
+                                AnnData/Scanpy
                                 (.h5ad) file</label>
                         </div>
                     </div>

main.js

@@ -71,7 +71,8 @@ async function predict(worker, modelID, h5File, cellRangePercent) {
         )} minutes`;
         resolve([cellNames, classes, predictions, coordinates]);
       } else if (type === "error") {
-        reject(error);
+        document.getElementById("message").textContent = error.message;
+        // reject(error);
       }
     };
 
@@ -310,9 +311,6 @@ async function main() {
           worker = new Worker("worker.js");
         }
 
-        // Reset the attention accumulator
-        worker.postMessage({ type: "resetAttentionAccumulator" });
-
         const [cellNames, classes, predictions, coordinates] = await predict(
           worker,
           modelID,

worker.js

@@ -4,25 +4,21 @@ self.importScripts(
   "https://cdn.jsdelivr.net/npm/[email protected]/lib/umap-js.min.js"
 );
 
-// Global variables for accumulators
-let attentionAccumulator = null;
+// Global variables
+self.model = null;
+self.attentionAccumulator = null;
 
 self.addEventListener("message", async function (event) {
   const { type } = event.data;
 
   if (type === "startPrediction") {
     predict(event);
-    // Handle prediction as shown above
-    // ...
   } else if (type === "getAttentionAccumulator") {
     self.postMessage({
       type: "attentionAccumulator",
-      attentionAccumulator: attentionAccumulator.buffer,
+      attentionAccumulator: self.attentionAccumulator.buffer,
       genes: self.model.genes,
     });
-  } else if (type === "resetAttentionAccumulator") {
-    attentionAccumulator = null;
-    self.postMessage({ type: "attentionAccumulatorReset" });
   }
 });
 
@@ -97,19 +93,16 @@ async function instantiateModel(id) {
 
   if (location.hostname === "localhost") {
     ort.env.debug = true;
-    ort.env.logLevel = "verbose";
-    ort.env.trace = true;
-    options["logSeverityLevel"] = 0;
-    options["logVerbosityLevel"] = 0;
+    // ort.env.logLevel = "verbose";
+    // ort.env.trace = true;
+    // options["logSeverityLevel"] = 0;
+    // options["logVerbosityLevel"] = 0;
   }
 
   // Create the InferenceSession with the model ArrayBuffer
   const session = await ort.InferenceSession.create(modelArray.buffer, options);
   console.log("Model Output names", session.outputNames);
 
-  // Initialize attention accumulator
-  attentionAccumulator = new Float32Array(genes.length);
-
   return { id, session, genes, classes };
 }
 
@@ -164,6 +157,9 @@ async function predict(event) {
       self.model = await instantiateModel(event.data.modelID);
     }
 
+    // Reset attention accumulator
+    self.attentionAccumulator = new Float32Array(self.model.genes.length);
+
     self.postMessage({ type: "status", message: "Loading file" });
     if (!FS.analyzePath("/work").exists) {
       FS.mkdir("/work");
@@ -250,12 +246,8 @@ async function predict(event) {
 
       encodings.push(output.encoding.cpuData);
 
-      if (!attentionAccumulator) {
-        attentionAccumulator = new Float32Array(genes.length);
-      }
-
-      for (let i = 0; i < attentionAccumulator.length; i++) {
-        attentionAccumulator[i] += output.attention.cpuData[i];
+      for (let i = 0; i < self.attentionAccumulator.length; i++) {
+        self.attentionAccumulator[i] += output.attention.cpuData[i];
       }
 
       // Post progress update
@@ -273,15 +265,22 @@ async function predict(event) {
       nEpochs: 400,
       nNeighbors: 15,
     });
-    const coordinates = await umap.fitAsync(encodings, (epochNumber) => {
-      // check progress and give user feedback, or return `false` to stop
-      self.postMessage({
-        type: "progress",
-        message: "Computing coordinates...",
-        countFinished: epochNumber,
-        totalToProcess: umap.getNEpochs(),
+
+    let coordinates = null;
+    try {
+      coordinates = await umap.fitAsync(encodings, (epochNumber) => {
+        // check progress and give user feedback, or return `false` to stop
+        self.postMessage({
+          type: "progress",
+          message: "Computing coordinates...",
+          countFinished: epochNumber,
+          totalToProcess: umap.getNEpochs(),
+        });
       });
-    });
+    } catch (error) {
+      self.postMessage({ type: "error", error });
+      throw error;
+    }
 
     annData.close();
     FS.unmount("/work");
@@ -301,6 +300,6 @@ async function predict(event) {
     });
   } catch (error) {
     FS.unmount("/work");
-    self.postMessage({ type: "error", error: error.message });
+    self.postMessage({ type: "error", error: error });
   }
 }