Add caching in Python

Still need it in JS

app/src/components/dashboard/LoggedCallTable.tsx

@@ -99,6 +99,14 @@ function TableRow({
     [loggedCall.tags],
   );
 
+  const durationCell = (
+    <Td isNumeric>
+      {loggedCall.cacheHit
+        ? "Cache hit"
+        : ((loggedCall.modelResponse?.durationMs ?? 0) / 1000).toFixed(2) + "s"}
+    </Td>
+  );
+
   return (
     <>
       <Tr
@@ -120,7 +128,7 @@ function TableRow({
           </Tooltip>
         </Td>
         <Td width="100%">{model}</Td>
-        <Td isNumeric>{((loggedCall.modelResponse?.durationMs ?? 0) / 1000).toFixed(2)}s</Td>
+        {durationCell}
         <Td isNumeric>{loggedCall.modelResponse?.inputTokens}</Td>
         <Td isNumeric>{loggedCall.modelResponse?.outputTokens}</Td>
         <Td sx={{ color: isError ? "red.500" : "green.500", fontWeight: "semibold" }} isNumeric>

app/src/server/api/routers/externalApi.router.ts

@@ -70,15 +70,9 @@ export const externalApiRouter = createTRPCRouter({
       const cacheKey = hashRequest(key.projectId, reqPayload as JsonValue);
 
       const existingResponse = await prisma.loggedCallModelResponse.findFirst({
-        where: {
-          cacheKey,
-        },
-        include: {
-          originalLoggedCall: true,
-        },
-        orderBy: {
-          requestedAt: "desc",
-        },
+        where: { cacheKey },
+        include: { originalLoggedCall: true },
+        orderBy: { requestedAt: "desc" },
       });
 
       if (!existingResponse) return { respPayload: null };