feat: update compression response to new API format

README.md

@@ -20,7 +20,11 @@ const edgee = new Edgee("your-api-key");
 
 // Send a simple request
 const response = await edgee.send({
+<<<<<<< HEAD
   model: 'gpt-5.2',
+=======
+  model: 'anthropic/claude-haiku-4-5',
+>>>>>>> ecec4c8 (feat: update compression response to new API format)
   input: 'What is the capital of France?',
 });
 
@@ -34,7 +38,7 @@ The `send()` method makes non-streaming chat completion requests:
 
 ```typescript
 const response = await edgee.send({
-  model: 'gpt-5.2',
+  model: 'anthropic/claude-haiku-4-5',
   input: 'Hello, world!',
 });
 
@@ -49,9 +53,10 @@ if (response.usage) {
 }
 
 if (response.compression) {
-  console.log(`Input tokens: ${response.compression.input_tokens}`);
   console.log(`Saved tokens: ${response.compression.saved_tokens}`);
-  console.log(`Compression rate: ${response.compression.rate}`);
+  console.log(`Reduction: ${response.compression.reduction}%`);
+  console.log(`Cost savings: $${(response.compression.cost_savings / 1_000_000).toFixed(3)}`);
+  console.log(`Time: ${response.compression.time_ms} ms`);
 }
 ```
 
@@ -60,7 +65,7 @@ if (response.compression) {
 The `stream()` method enables real-time streaming responses:
 
 ```typescript
-for await (const chunk of edgee.stream('gpt-5.2', 'Tell me a story')) {
+for await (const chunk of edgee.stream('anthropic/claude-haiku-4-5', 'Tell me a story')) {
   if (chunk.text) {
     process.stdout.write(chunk.text);
   }

example/compression.ts

@@ -81,7 +81,7 @@ ${LARGE_CONTEXT}
 Based on this context, summarize the key milestones in AI development in 3 bullet points.`;
 
 const response = await edgee.send({
-  model: "gpt-5.2",
+  model: "anthropic/claude-haiku-4-5",
   input: {
     messages: [{ role: "user", content: userMessage }],
     enable_compression: true,
@@ -104,26 +104,25 @@ if (response.usage) {
 // Display compression information
 if (response.compression) {
   console.log("Compression Metrics:");
-  console.log(`  Input tokens:  ${response.compression.input_tokens}`);
   console.log(`  Saved tokens:  ${response.compression.saved_tokens}`);
+  console.log(`  Reduction:     ${response.compression.reduction}%`);
   console.log(
-    `  Compression rate: ${(response.compression.rate * 100).toFixed(2)}%`
-  );
-  const savingsPct =
-    response.compression.input_tokens > 0
-      ? (response.compression.saved_tokens /
-          response.compression.input_tokens) *
-        100
-      : 0;
-  console.log(`  Savings: ${savingsPct.toFixed(1)}% of input tokens saved!`);
-  console.log();
-  console.log(`  💡 Without compression, this request would have used`);
-  console.log(`     ${response.compression.input_tokens} input tokens.`);
-  console.log(
-    `     With compression, only ${
-      response.compression.input_tokens - response.compression.saved_tokens
-    } tokens were processed!`
+    `  Cost savings:  $${(response.compression.cost_savings / 1_000_000).toFixed(3)}`
   );
+  console.log(`  Time:          ${response.compression.time_ms} ms`);
+  if (response.compression.reduction > 0) {
+    const originalTokens =
+      Math.floor(
+        (response.compression.saved_tokens * 100) / response.compression.reduction
+      );
+    const tokensAfter = originalTokens - response.compression.saved_tokens;
+    console.log();
+    console.log(`  💡 Without compression, this request would have used`);
+    console.log(`     ${originalTokens} input tokens.`);
+    console.log(
+      `     With compression, only ${tokensAfter} tokens were processed!`
+    );
+  }
 } else {
   console.log("No compression data available in response.");
   console.log(

example/package.json

@@ -3,7 +3,8 @@
   "version": "1.0.0",
   "type": "module",
   "scripts": {
-    "test": "npx tsx test.ts"
+    "test": "npx tsx test.ts",
+    "test:compression": "npx tsx compression.ts"
   },
   "dependencies": {
     "edgee": "file:.."

example/test.ts

@@ -5,7 +5,7 @@ const edgee = new Edgee(process.env.EDGEE_API_KEY || "test-key");
 // Test 1: Simple string input
 console.log("Test 1: Simple string input");
 const response1 = await edgee.send({
-  model: "gpt-5.2",
+  model: "anthropic/claude-haiku-4-5",
   input: "What is the capital of France?",
 });
 console.log("Content:", response1.choices[0].message.content);
@@ -15,7 +15,7 @@ console.log();
 // Test 2: Full input object with messages
 console.log("Test 2: Full input object with messages");
 const response2 = await edgee.send({
-  model: "gpt-5.2",
+  model: "anthropic/claude-haiku-4-5",
   input: {
     messages: [
       { role: "system", content: "You are a helpful assistant." },
@@ -29,7 +29,7 @@ console.log();
 // Test 3: With tools
 console.log("Test 3: With tools");
 const response3 = await edgee.send({
-  model: "gpt-5.2",
+  model: "anthropic/claude-haiku-4-5",
   input: {
     messages: [{ role: "user", content: "What is the weather in Paris?" }],
     tools: [
@@ -56,3 +56,16 @@ console.log(
   "Tool calls:",
   JSON.stringify(response3.choices[0].message.tool_calls, null, 2)
 );
+console.log();
+
+// Test 4: Streaming
+console.log("Test 4: Streaming");
+for await (const chunk of edgee.stream(
+  "anthropic/claude-haiku-4-5",
+  "What is Typescript?"
+)) {
+  if (chunk.text) {
+    process.stdout.write(chunk.text);
+  }
+}
+console.log("\n");