feat: update compression response to new API format, add SSL cert fix

README.md

@@ -20,7 +20,7 @@ edgee = Edgee("your-api-key")
 
 # Send a simple request
 response = edgee.send(
-    model="gpt-5.2",
+    model="anthropic/claude-haiku-4-5",
     input="What is the capital of France?"
 )
 
@@ -34,7 +34,7 @@ The `send()` method makes non-streaming chat completion requests:
 
 ```python
 response = edgee.send(
-    model="gpt-5.2",
+    model="anthropic/claude-haiku-4-5",
     input="Hello, world!"
 )
 
@@ -48,17 +48,18 @@ if response.usage:
     print(f"Tokens used: {response.usage.total_tokens}")
 
 if response.compression:
-    print(f"Input tokens: {response.compression.input_tokens}")
     print(f"Saved tokens: {response.compression.saved_tokens}")
-    print(f"Compression rate: {response.compression.rate}")
+    print(f"Reduction: {response.compression.reduction}%")
+    print(f"Cost savings: ${response.compression.cost_savings / 1_000_000:.3f}")
+    print(f"Time: {response.compression.time_ms} ms")
 ```
 
 ## Stream Method
 
 The `stream()` method enables real-time streaming responses:
 
 ```python
-for chunk in edgee.stream("gpt-5.2", "Tell me a story"):
+for chunk in edgee.stream("anthropic/claude-haiku-4-5", "Tell me a story"):
     if chunk.text:
         print(chunk.text, end="", flush=True)
 

edgee/__init__.py

@@ -2,6 +2,7 @@
 
 import json
 import os
+import ssl
 from dataclasses import dataclass
 from urllib.error import HTTPError
 from urllib.request import Request, urlopen
@@ -11,6 +12,18 @@
 API_ENDPOINT = "/v1/chat/completions"
 
 
+def _ssl_context() -> ssl.SSLContext:
+    """Create SSL context. Uses certifi's CA bundle when available (fixes cert issues on macOS)."""
+    ctx = ssl.create_default_context()
+    try:
+        import certifi
+
+        ctx.load_verify_locations(certifi.where())
+    except ImportError:
+        pass  # Use default system/store certs
+    return ctx
+
+
 @dataclass
 class FunctionDefinition:
     name: str
@@ -70,9 +83,10 @@ class Usage:
 
 @dataclass
 class Compression:
-    input_tokens: int
     saved_tokens: int
-    rate: float
+    cost_savings: int  # micro-units (e.g. 27000 = $0.027)
+    reduction: int  # percentage (e.g. 48 = 48%)
+    time_ms: int  # milliseconds
 
 
 @dataclass
@@ -253,7 +267,7 @@ def send(
     def _handle_non_streaming_response(self, request: Request) -> SendResponse:
         """Handle non-streaming response."""
         try:
-            with urlopen(request) as response:
+            with urlopen(request, context=_ssl_context()) as response:
                 data = json.loads(response.read().decode("utf-8"))
         except HTTPError as e:
             error_body = e.read().decode("utf-8")
@@ -279,17 +293,18 @@ def _handle_non_streaming_response(self, request: Request) -> SendResponse:
         compression = None
         if "compression" in data:
             compression = Compression(
-                input_tokens=data["compression"]["input_tokens"],
                 saved_tokens=data["compression"]["saved_tokens"],
-                rate=data["compression"]["rate"],
+                cost_savings=data["compression"]["cost_savings"],
+                reduction=data["compression"]["reduction"],
+                time_ms=data["compression"]["time_ms"],
             )
 
         return SendResponse(choices=choices, usage=usage, compression=compression)
 
     def _handle_streaming_response(self, request: Request):
         """Handle streaming response, yielding StreamChunk objects."""
         try:
-            with urlopen(request) as response:
+            with urlopen(request, context=_ssl_context()) as response:
                 # Read and parse SSE stream
                 for line in response:
                     decoded_line = line.decode("utf-8")

example/compression.py

@@ -87,7 +87,7 @@
 Based on this context, summarize the key milestones in AI development in 3 bullet points."""
 
 response = edgee.send(
-    model="gpt-5.2",
+    model="anthropic/claude-haiku-4-5",
     input={
         "messages": [
             {"role": "user", "content": user_message},
@@ -111,21 +111,17 @@
 # Display compression information
 if response.compression:
     print("Compression Metrics:")
-    print(f"  Input tokens:  {response.compression.input_tokens}")
     print(f"  Saved tokens:  {response.compression.saved_tokens}")
-    print(f"  Compression rate: {response.compression.rate:.2%}")
-    savings_pct = (
-        (response.compression.saved_tokens / response.compression.input_tokens * 100)
-        if response.compression.input_tokens > 0
-        else 0
-    )
-    print(f"  Savings: {savings_pct:.1f}% of input tokens saved!")
-    print()
-    print("  💡 Without compression, this request would have used")
-    print(f"     {response.compression.input_tokens} input tokens.")
-    print(
-        f"     With compression, only {response.compression.input_tokens - response.compression.saved_tokens} tokens were processed!"
-    )
+    print(f"  Reduction:     {response.compression.reduction}%")
+    print(f"  Cost savings:  ${response.compression.cost_savings / 1_000_000:.3f}")
+    print(f"  Time:          {response.compression.time_ms} ms")
+    if response.compression.reduction > 0:
+        original_tokens = response.compression.saved_tokens * 100 // response.compression.reduction
+        tokens_after = original_tokens - response.compression.saved_tokens
+        print()
+        print("  💡 Without compression, this request would have used")
+        print(f"     {original_tokens} input tokens.")
+        print(f"     With compression, only {tokens_after} tokens were processed!")
 else:
     print("No compression data available in response.")
     print("Note: Compression data is only returned when compression is enabled")

example/test.py

@@ -13,7 +13,7 @@
 # Test 1: Simple string input
 print("Test 1: Simple string input")
 response1 = edgee.send(
-    model="mistral/mistral-small-latest",
+    model="anthropic/claude-haiku-4-5",
     input="What is the capital of France?",
 )
 print(f"Content: {response1.text}")
@@ -23,7 +23,7 @@
 # Test 2: Full input object with messages
 print("Test 2: Full input object with messages")
 response2 = edgee.send(
-    model="mistral/mistral-small-latest",
+    model="anthropic/claude-haiku-4-5",
     input={
         "messages": [
             {"role": "system", "content": "You are a helpful assistant."},
@@ -37,7 +37,7 @@
 # Test 3: With tools
 print("Test 3: With tools")
 response3 = edgee.send(
-    model="gpt-5.2",
+    model="anthropic/claude-haiku-4-5",
     input={
         "messages": [{"role": "user", "content": "What is the weather in Paris?"}],
         "tools": [
@@ -65,7 +65,7 @@
 
 # Test 4: Streaming
 print("Test 4: Streaming")
-for chunk in edgee.stream(model="mistral/mistral-small-latest", input="What is Python?"):
+for chunk in edgee.stream(model="anthropic/claude-haiku-4-5", input="What is Python?"):
     if chunk.text:
         print(chunk.text, end="", flush=True)
 print("\n")

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "edgee"
-version = "1.0.1"
+version = "1.0.2"
 description = "Lightweight Python SDK for Edgee AI Gateway"
 readme = "README.md"
 license = "Apache-2.0"
@@ -25,6 +25,8 @@ Repository = "https://github.com/edgee-ai/python-sdk"
 
 [project.optional-dependencies]
 dev = ["pytest>=8.0.0", "ruff>=0.8.0"]
+# Install certifi to fix SSL certificate verification on macOS (Python.org installs)
+ssl = ["certifi>=2024.0.0"]
 
 [build-system]
 requires = ["hatchling"]

tests/test_edgee.py

@@ -320,9 +320,10 @@ def test_send_with_compression_response(self, mock_urlopen):
             ],
             "usage": {"prompt_tokens": 100, "completion_tokens": 50, "total_tokens": 150},
             "compression": {
-                "input_tokens": 100,
                 "saved_tokens": 42,
-                "rate": 0.6102003642987249,
+                "cost_savings": 27000,
+                "reduction": 48,
+                "time_ms": 150,
             },
         }
         mock_urlopen.return_value = self._mock_response(mock_response_data)
@@ -331,9 +332,10 @@ def test_send_with_compression_response(self, mock_urlopen):
         result = client.send(model="gpt-4", input="Test")
 
         assert result.compression is not None
-        assert result.compression.input_tokens == 100
         assert result.compression.saved_tokens == 42
-        assert result.compression.rate == 0.6102003642987249
+        assert result.compression.cost_savings == 27000
+        assert result.compression.reduction == 48
+        assert result.compression.time_ms == 150
 
     @patch("edgee.urlopen")
     def test_send_without_compression_response(self, mock_urlopen):