confamnode 0.2.5__tar.gz → 0.2.6__tar.gz

{confamnode-0.2.5 → confamnode-0.2.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: confamnode
-Version: 0.2.5
+Version: 0.2.6
 Summary: The Nigerian AI inference gateway
 Project-URL: Repository, https://github.com/confamnodeai/confamnode
 Project-URL: Bug Tracker, https://github.com/confamnodeai/confamnode/issues
@@ -258,7 +258,7 @@ ansa = client.gist(
 
 Caching is controlled **per request** and is **off by default** — every call returns a fresh response, even when the request is identical. This keeps data-generation loops and any workflow that resends the same prompt from getting the same cached answer back each time.
 
-Pass `cache=True` to read from and write to the cache — useful for idempotent lookups or to save cost on repeated queries:
+Pass `cache=True` to let the gateway serve and store a cached response — useful for idempotent lookups:
 
 ```python
 # Default — caching off, fresh response every call
@@ -267,8 +267,8 @@ ansa = client.gist(
     messages="How you dey?"
 )
 
-# Enable caching — use a stored response when the request matches,
-# and store this response for next time
+# Enable caching — the gateway may return a stored response for a
+# matching request, and store this one for next time
 ansa = client.gist(
     model="confam-speed",
     messages="How you dey?",
@@ -276,7 +276,7 @@ ansa = client.gist(
 )
 ```
 
-A cache hit is typically returned near-instantly and at little or no token cost — a quick way to confirm caching is active.
+Caching must be enabled for your account for `cache=True` to take effect. A cache hit returns near-instantly — the quickest way to see it is to send the **same** request twice with `cache=True`: the first call populates the cache, the second is served from it.
 
 ---
 

{confamnode-0.2.5 → confamnode-0.2.6}/README.md

@@ -234,7 +234,7 @@ ansa = client.gist(
 
 Caching is controlled **per request** and is **off by default** — every call returns a fresh response, even when the request is identical. This keeps data-generation loops and any workflow that resends the same prompt from getting the same cached answer back each time.
 
-Pass `cache=True` to read from and write to the cache — useful for idempotent lookups or to save cost on repeated queries:
+Pass `cache=True` to let the gateway serve and store a cached response — useful for idempotent lookups:
 
 ```python
 # Default — caching off, fresh response every call
@@ -243,8 +243,8 @@ ansa = client.gist(
     messages="How you dey?"
 )
 
-# Enable caching — use a stored response when the request matches,
-# and store this response for next time
+# Enable caching — the gateway may return a stored response for a
+# matching request, and store this one for next time
 ansa = client.gist(
     model="confam-speed",
     messages="How you dey?",
@@ -252,7 +252,7 @@ ansa = client.gist(
 )
 ```
 
-A cache hit is typically returned near-instantly and at little or no token cost — a quick way to confirm caching is active.
+Caching must be enabled for your account for `cache=True` to take effect. A cache hit returns near-instantly — the quickest way to see it is to send the **same** request twice with `cache=True`: the first call populates the cache, the second is served from it.
 
 ---
 

{confamnode-0.2.5 → confamnode-0.2.6}/confamnode/__init__.py

@@ -8,7 +8,7 @@ from confamnode.exceptions import (
 from confamnode.ansa import Ansa, Usage, Cost
 from confamnode import models
 
-__version__ = "0.2.5"
+__version__ = "0.2.6"
 
 __all__ = [
     "ConfamNode",

{confamnode-0.2.5 → confamnode-0.2.6}/confamnode/client.py

@@ -61,20 +61,17 @@ class ConfamNode:
             else:
                 body["system"] = system # None or custom string
 
-        # Caching is controlled per request. By default the SDK opts OUT, so
-        # identical requests (same model + messages + system) each return a
+        # Caching is opt-in. By default the SDK sends an explicit cache-bypass,
+        # so identical requests (same model + messages + system) each return a
         # FRESH response -- important for data-generation loops that resend the
-        # same prompt expecting varied output. Pass cache=True to read from and
-        # write to the cache (idempotent lookups, cost savings on repeats).
+        # same prompt expecting varied output.
         #
-        # These flags only tell the gateway whether to use a cache for THIS
-        # request; whether one exists at all is a gateway-side capability.
-        if cache:
-            body["cache"] = {
-                "no-cache": False, # Cache the response
-                "no-store": False # Store the response
-            }
-        else:
+        # With cache=True we send NO cache field at all, letting the gateway
+        # apply its normal caching. That absent-field shape is exactly what
+        # produced cached responses before this flag existed, so it's the
+        # request form known to engage the cache -- more reliable than sending
+        # explicit "false" flags the gateway may not interpret identically.
+        if not cache:
             body["cache"] = {
                 "no-cache": True, # Skip cache check, get fresh response
                 "no-store": True # Don't cache this response

{confamnode-0.2.5 → confamnode-0.2.6}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "confamnode"
-version = "0.2.5"
+version = "0.2.6"
 description = "The Nigerian AI inference gateway"
 readme = "README.md"
 requires-python = ">=3.10"

{confamnode-0.2.5 → confamnode-0.2.6}/tests/test_cache_flag.py

@@ -18,7 +18,6 @@ API_KEY = "confam-test"
 MODEL = next(iter(VALID_MODELS))
 
 CACHE_OFF = {"no-cache": True, "no-store": True}    # skip read + skip store
-CACHE_ON = {"no-cache": False, "no-store": False}   # use + store
 
 
 @pytest.fixture
@@ -55,20 +54,13 @@ def test_cache_off_by_default(captured_body):
     assert captured_body["body"]["cache"] == CACHE_OFF
 
 
-def test_cache_true_uses_and_stores(captured_body):
+def test_cache_true_omits_field(captured_body):
+    # cache=True lets the gateway apply its normal caching by sending NO cache
+    # field -- the request shape proven to engage the cache.
     _gist(cache=True)
-    assert captured_body["body"]["cache"] == CACHE_ON
+    assert "cache" not in captured_body["body"]
 
 
 def test_cache_false_is_explicit_off(captured_body):
     _gist(cache=False)
-    assert captured_body["body"]["cache"] == CACHE_OFF
-
-
-def test_cache_field_always_present(captured_body):
-    # The request should always state its caching intent in both directions,
-    # never leave it implied by omission.
-    _gist()
-    assert "cache" in captured_body["body"]
-    _gist(cache=True)
-    assert "cache" in captured_body["body"]
+    assert captured_body["body"]["cache"] == CACHE_OFF

{confamnode-0.2.5 → confamnode-0.2.6}/uv.lock

@@ -40,7 +40,7 @@ wheels = [
 
 [[package]]
 name = "confamnode"
-version = "0.2.4"
+version = "0.2.6"
 source = { editable = "." }
 dependencies = [
     { name = "httpx" },