python-fastllm 0.0.56__py3-none-any.whl → 0.0.58__py3-none-any.whl

fastllm/__init__.py

@@ -2,4 +2,4 @@
 
 - `fastllm.responses`: Provider-independent Responses resources and streaming over FastLLM"""
 
-__version__ = "0.0.56"
+__version__ = "0.0.58"

fastllm/acomplete.py

@@ -39,7 +39,7 @@ vendor_mapping = {
     "fireworks_ai": ('openai_chat', "https://api.fireworks.ai/inference/v1", "FIREWORKS_API_KEY"),
     "qwen":         ('openai_chat', "https://dashscope.aliyuncs.com/compatible-mode/v1", "QWEN_API_KEY"),
     "minimax":      ('anthropic', "https://api.minimax.io/anthropic", "MINIMAX_API_KEY"),
-    "meta":         ('openai', "https://api.meta.ai/v1", "META_API_KEY"),
+    "meta":         ('openai_chat', "https://api.meta.ai/v1", "META_API_KEY"),
     "zai":          ('openai_chat', "https://api.z.ai/api/paas/v4", "ZAI_API_KEY")}
 
 # %% ../nbs/06_acomplete.ipynb #e3ed40fb

fastllm/anthropic.py

@@ -44,7 +44,8 @@ def norm_usage(resp):
     pt = int(usg.get("input_tokens", 0) or 0) + cached + cache_creation
     ct = int(usg.get("output_tokens", 0) or 0)
     return Usage(prompt_tokens=pt, completion_tokens=ct, total_tokens=pt + ct,
-                 cached_tokens=cached, cache_creation_tokens=cache_creation, reasoning_tokens=0, raw=usg)
+                 cached_tokens=cached, cache_creation_tokens=cache_creation, reasoning_tokens=0, raw=usg,
+                 search_queries=int(nested_idx(usg, 'server_tool_use', 'web_search_requests') or 0))
 
 def finalize_usage(usg, parts):
     "Split reasoning tokens out of `output_tokens`: the reported `thinking_tokens`, else an estimate from the thinking text."
@@ -53,7 +54,7 @@ def finalize_usage(usg, parts):
     ct = int(usg.raw.get('output_tokens', usg.completion_tokens) or 0)
     rt = int(nested_idx(usg.raw, 'output_tokens_details', 'thinking_tokens') or 0) or (min(int(len(rc.split())*1.5), ct) if rc else 0)
     return Usage(prompt_tokens=usg.prompt_tokens, completion_tokens=ct-rt, total_tokens=usg.prompt_tokens+ct,
-        cached_tokens=usg.cached_tokens, cache_creation_tokens=usg.cache_creation_tokens, reasoning_tokens=rt, raw=usg.raw)
+        cached_tokens=usg.cached_tokens, cache_creation_tokens=usg.cache_creation_tokens, reasoning_tokens=rt, search_queries=usg.search_queries, raw=usg.raw)
 
 # %% ../nbs/04_anthropic.ipynb #7a8b1f8f
 def norm_finish(resp, tcs=None):

fastllm/gemini.py

@@ -67,7 +67,7 @@ def norm_usage(resp):
     if call := grounding(cand)[1]: stu['google_search'] = len(call.arguments['queries'] or [1])
     if stu: usg['server_tool_use'] = stu
     return Usage(prompt_tokens=pt, completion_tokens=ct, total_tokens=tt,
-                 cached_tokens=cached, reasoning_tokens=reasoning, raw=usg)
+                 cached_tokens=cached, reasoning_tokens=reasoning, search_queries=stu.get('google_search', 0), raw=usg)
 
 # %% ../nbs/05_gemini.ipynb #7a8b1f8f
 def norm_finish(resp, tcs=None):

fastllm/openai_responses.py

@@ -53,8 +53,7 @@ def norm_usage(resp):
     server_tool_use = dict(Counter(o['type'] for o in resp.get('output', []) if o.get('type') != 'function_call' and o.get('type', '').endswith('_call')))
     if server_tool_use: usg['server_tool_use'] = server_tool_use
     return Usage(prompt_tokens=pt, completion_tokens=ct, total_tokens=tt,
-        cached_tokens=cached, reasoning_tokens=reasoning, raw=usg)
-
+        cached_tokens=cached, reasoning_tokens=reasoning, search_queries=server_tool_use.get('web_search_call', 0), raw=usg)
 
 # %% ../nbs/02_oai_responses.ipynb #95102df4
 def norm_finish(resp, tcs=None):

fastllm/types.py

@@ -26,16 +26,15 @@ from aidialog.msg_parts import Msg, Thinking, ToolUse, Completion, StopResponse,
 class Usage(BasicRepr):
     "Normalized usage."
     def __init__(self, prompt_tokens=0, completion_tokens=0, total_tokens=0, cached_tokens=0,
-        cache_creation_tokens=0, reasoning_tokens=0, raw=None):
+        cache_creation_tokens=0, reasoning_tokens=0, search_queries=0, raw=None):
         if raw is None: raw = {}
         store_attr()
     def __eq__(self, o): return type(o) is type(self) and self.__dict__ == o.__dict__
 
-
 # %% ../nbs/00_types.ipynb #bae1c130
 @patch
 def __add__(self:Usage, o):
-    fields = 'prompt_tokens completion_tokens total_tokens cached_tokens cache_creation_tokens reasoning_tokens'.split()
+    fields = 'prompt_tokens completion_tokens total_tokens cached_tokens cache_creation_tokens reasoning_tokens search_queries'.split()
     return Usage(**{k: getattr(self, k)+getattr(o, k) for k in fields}, raw=o.raw)
 
 # %% ../nbs/00_types.ipynb #4901e693
@@ -366,4 +365,5 @@ def cost(self:Completion):
     api = api_registry[self.api_name]
     if not hasattr(api, 'cost'): raise NotImplementedError(f"API: {self.api_name} doesn't have a registered `cost` function in ns")
     res = api.cost(self.usage, meta)
+    if (n := self.usage.search_queries) and (rates := meta.get('search_context_cost_per_query')): res += n * rates['search_context_size_medium']
     return res*2 if self.vendor_name=='deepseek' and self.model.startswith('deepseek-v4') and is_deepseek_peak_hour() else res

{python_fastllm-0.0.56.dist-info → python_fastllm-0.0.58.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: python-fastllm
-Version: 0.0.56
+Version: 0.0.58
 Author-email: Kerem Turgutlu <keremturgutlu@gmail.com>
 License: Apache-2.0
 Project-URL: Repository, https://github.com/AnswerDotAI/fastllm

python_fastllm-0.0.58.dist-info/RECORD

@@ -0,0 +1,18 @@
+fastllm/__init__.py,sha256=uDoIIuBjxZKcbyRrdFaaGJlfJwWkr9dF7dNA3b04A9k,131
+fastllm/_modidx.py,sha256=qzUtLbjOche-GNbeNun3VRIQdcm5g4dUOCLJ6T9fG-w,46374
+fastllm/acomplete.py,sha256=CusLue35xiaswpmTZyLkJj6t4CVsNDeeY3Sk7hzVNoY,10777
+fastllm/anthropic.py,sha256=ktH-EuQMQnp8O2_6VmSrvvDCIe2QTeSR7jNpWfqf3L4,18884
+fastllm/chat.py,sha256=klji3SOYs5xyzrSFbNhy3uoQ2TSZ3a-8HTjvdRtaG5Q,25036
+fastllm/codex.py,sha256=r6auZ8egia83aGkWcF4132AzW4SsBvKeDwg_CxyXj9E,5808
+fastllm/gemini.py,sha256=vpY_f4Jqn344l9CANm6_c7Cmhjq2URXxWhdVv_LswSg,17615
+fastllm/model_prices.json,sha256=pa0j96LZgklYii8howX443QbukUP1iFr3lBcucS-6Yo,2133018
+fastllm/openai_chat.py,sha256=n5INfDCsAZIiKUImWeaX6SJ_OAJ_uMhw20HUkgFgdgM,12379
+fastllm/openai_responses.py,sha256=NrvzNCKkSQabbAMnxYCSPdROaT1u8Vp2eoRzbzAA-4M,15345
+fastllm/responses.py,sha256=9gh-UlhfswCcP0RCzLuPT5meWNh9mF_OIivzCzZU94I,22726
+fastllm/streaming.py,sha256=DrqNJkMhyntFKokY6GF6oW3CjELrk5XWmnxwcApJUyM,8770
+fastllm/types.py,sha256=OznlKQZUx2DrLE_O-sbIBsT6q6EBeVDlH9kXxhsGBSE,19445
+python_fastllm-0.0.58.dist-info/METADATA,sha256=scQa4XA1Oh28m6HO6knQtPmOwV_E0ppqFdXuMH8GTA8,18537
+python_fastllm-0.0.58.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
+python_fastllm-0.0.58.dist-info/entry_points.txt,sha256=WkWd04F3xVYFP5bsexaQe9Se6uelwE5ULoTtHWreO0s,133
+python_fastllm-0.0.58.dist-info/top_level.txt,sha256=F8qodL7nEGUHGmzzqfhNKCTIr1i0D6cvudOnm3z7o0Y,8
+python_fastllm-0.0.58.dist-info/RECORD,,

python_fastllm-0.0.56.dist-info/RECORD

@@ -1,18 +0,0 @@
-fastllm/__init__.py,sha256=OWNw1jlF14M2QUzFV9voeL-APa-sSFSfenIVDvWhGFQ,131
-fastllm/_modidx.py,sha256=qzUtLbjOche-GNbeNun3VRIQdcm5g4dUOCLJ6T9fG-w,46374
-fastllm/acomplete.py,sha256=64-k3LgLoX1EoWcHBS_-epZ-JvX5JL-As9QTo-WSzDc,10772
-fastllm/anthropic.py,sha256=vsWkmIxIh8j7LTPfLzgrNggOl_2LuRKKcdVhLfDDdRE,18748
-fastllm/chat.py,sha256=klji3SOYs5xyzrSFbNhy3uoQ2TSZ3a-8HTjvdRtaG5Q,25036
-fastllm/codex.py,sha256=r6auZ8egia83aGkWcF4132AzW4SsBvKeDwg_CxyXj9E,5808
-fastllm/gemini.py,sha256=978x7WDug5nSNvTOacap3bglgQhddahykFQvoOKOvhs,17571
-fastllm/model_prices.json,sha256=pa0j96LZgklYii8howX443QbukUP1iFr3lBcucS-6Yo,2133018
-fastllm/openai_chat.py,sha256=n5INfDCsAZIiKUImWeaX6SJ_OAJ_uMhw20HUkgFgdgM,12379
-fastllm/openai_responses.py,sha256=O30JBRG0kTI31FPnzAOMEE8iiIsMBdAxVGoZys58Jn4,15288
-fastllm/responses.py,sha256=9gh-UlhfswCcP0RCzLuPT5meWNh9mF_OIivzCzZU94I,22726
-fastllm/streaming.py,sha256=DrqNJkMhyntFKokY6GF6oW3CjELrk5XWmnxwcApJUyM,8770
-fastllm/types.py,sha256=hET6SNjqC8Z5PasrZ4uU-OObIfs3QyI0ZPY4jE2rh1Q,19268
-python_fastllm-0.0.56.dist-info/METADATA,sha256=NhRX9Fj4A7nRSg-bjmytouLTvjaE5TIYVLnZc3prEvM,18537
-python_fastllm-0.0.56.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
-python_fastllm-0.0.56.dist-info/entry_points.txt,sha256=WkWd04F3xVYFP5bsexaQe9Se6uelwE5ULoTtHWreO0s,133
-python_fastllm-0.0.56.dist-info/top_level.txt,sha256=F8qodL7nEGUHGmzzqfhNKCTIr1i0D6cvudOnm3z7o0Y,8
-python_fastllm-0.0.56.dist-info/RECORD,,