not-again-ai 0.10.3__py3-none-any.whl → 0.11.0__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (9) hide show
  1. not_again_ai/llm/openai_api/chat_completion.py +12 -12
  2. not_again_ai/local_llm/chat_completion.py +11 -0
  3. not_again_ai/local_llm/ollama/chat_completion.py +19 -4
  4. not_again_ai/local_llm/ollama/model_mapping.py +1 -0
  5. {not_again_ai-0.10.3.dist-info → not_again_ai-0.11.0.dist-info}/METADATA +5 -5
  6. {not_again_ai-0.10.3.dist-info → not_again_ai-0.11.0.dist-info}/RECORD +9 -9
  7. {not_again_ai-0.10.3.dist-info → not_again_ai-0.11.0.dist-info}/LICENSE +0 -0
  8. {not_again_ai-0.10.3.dist-info → not_again_ai-0.11.0.dist-info}/WHEEL +0 -0
  9. {not_again_ai-0.10.3.dist-info → not_again_ai-0.11.0.dist-info}/entry_points.txt +0 -0
not_again_ai/llm/openai_api/chat_completion.py CHANGED
@@ -32,8 +32,8 @@ def chat_completion(
32
32
  https://platform.openai.com/docs/models/model-endpoint-compatibility
33
33
  for details on which models work with the Chat API.
34
34
  client (OpenAI): An instance of the OpenAI client.
35
- tools (list[dict[str, Any]], optional): A list of tools the model may generate JSON inputs for.
36
- Defaults to None.
35
+ tools (list[dict[str, Any]], optional):A list of tools the model may call.
36
+ Use this to provide a list of functions the model may generate JSON inputs for. Defaults to None.
37
37
  tool_choice (str, optional): The tool choice to use. Can be "auto", "required", "none", or a specific function name.
38
38
  Note the function name cannot be any of "auto", "required", or "none". Defaults to "auto".
39
39
  max_tokens (int, optional): The maximum number of tokens to generate in the chat completion.
@@ -58,21 +58,21 @@ def chat_completion(
58
58
 
59
59
  Returns:
60
60
  dict[str, Any]: A dictionary with the following keys:
61
- 'finish_reason' (str): The reason the model stopped generating further tokens.
61
+ finish_reason (str): The reason the model stopped generating further tokens.
62
62
  Can be 'stop', 'length', or 'tool_calls'.
63
- 'tool_names' (list[str], optional): The names of the tools called by the model.
64
- 'tool_args_list' (list[dict], optional): The arguments of the tools called by the model.
65
- 'message' (str | dict): The content of the generated assistant message.
63
+ tool_names (list[str], optional): The names of the tools called by the model.
64
+ tool_args_list (list[dict], optional): The arguments of the tools called by the model.
65
+ message (str | dict): The content of the generated assistant message.
66
66
  If json_mode is True, this will be a dictionary.
67
- 'logprobs' (list[dict[str, Any] | list[dict[str, Any]]]): If logprobs[1] is between 1 and 5, each element in the list
67
+ logprobs (list[dict[str, Any] | list[dict[str, Any]]]): If logprobs[1] is between 1 and 5, each element in the list
68
68
  will be a list of dictionaries containing the token, logprob, and bytes for the top `logprobs[1]` logprobs. Otherwise,
69
69
  this will be a list of dictionaries containing the token, logprob, and bytes for each token in the message.
70
- 'choices' (list[dict], optional): A list of chat completion choices if n > 1 where each dict contains the above fields.
71
- 'completion_tokens' (int): The number of tokens used by the model to generate the completion.
70
+ choices (list[dict], optional): A list of chat completion choices if n > 1 where each dict contains the above fields.
71
+ completion_tokens (int): The number of tokens used by the model to generate the completion.
72
72
  NOTE: If n > 1 this is the sum of all completions.
73
- 'prompt_tokens' (int): The number of tokens in the messages sent to the model.
74
- 'system_fingerprint' (str, optional): If seed is set, a unique identifier for the model used to generate the response.
75
- 'response_duration' (float): The time, in seconds, taken to generate the response from the API.
73
+ prompt_tokens (int): The number of tokens in the messages sent to the model.
74
+ system_fingerprint (str, optional): If seed is set, a unique identifier for the model used to generate the response.
75
+ response_duration (float): The time, in seconds, taken to generate the response from the API.
76
76
  """
77
77
  response_format = {"type": "json_object"} if json_mode else None
78
78
 
not_again_ai/local_llm/chat_completion.py CHANGED
@@ -11,6 +11,7 @@ def chat_completion(
11
11
  messages: list[dict[str, Any]],
12
12
  model: str,
13
13
  client: OpenAI | Client,
14
+ tools: list[dict[str, Any]] | None = None,
14
15
  max_tokens: int | None = None,
15
16
  temperature: float = 0.7,
16
17
  json_mode: bool = False,
@@ -25,6 +26,8 @@ def chat_completion(
25
26
  messages (list[dict[str, Any]]): A list of messages to send to the model.
26
27
  model (str): The model name to use.
27
28
  client (OpenAI | Client): The client object to use for chat completion.
29
+ tools (list[dict[str, Any]], optional):A list of tools the model may call.
30
+ Use this to provide a list of functions the model may generate JSON inputs for. Defaults to None.
28
31
  max_tokens (int, optional): The maximum number of tokens to generate.
29
32
  temperature (float, optional): The temperature of the model. Increasing the temperature will make the model answer more creatively.
30
33
  json_mode (bool, optional): This will structure the response as a valid JSON object.
@@ -34,6 +37,9 @@ def chat_completion(
34
37
  dict[str, Any]: A dictionary with the following keys
35
38
  message (str | dict): The content of the generated assistant message.
36
39
  If json_mode is True, this will be a dictionary.
40
+ tool_names (list[str], optional): The names of the tools called by the model.
41
+ If the model does not support tools, a ResponseError is raised.
42
+ tool_args_list (list[dict], optional): The arguments of the tools called by the model.
37
43
  prompt_tokens (int): The number of tokens in the messages sent to the model.
38
44
  completion_tokens (int): The number of tokens used by the model to generate the completion.
39
45
  response_duration (float): The time, in seconds, taken to generate the response by using the model.
@@ -45,6 +51,7 @@ def chat_completion(
45
51
  messages=messages,
46
52
  model=model,
47
53
  client=client,
54
+ tools=tools,
48
55
  max_tokens=max_tokens,
49
56
  temperature=temperature,
50
57
  json_mode=json_mode,
@@ -56,6 +63,7 @@ def chat_completion(
56
63
  messages=messages,
57
64
  model=model,
58
65
  client=client,
66
+ tools=tools,
59
67
  max_tokens=max_tokens,
60
68
  temperature=temperature,
61
69
  json_mode=json_mode,
@@ -68,6 +76,9 @@ def chat_completion(
68
76
  # Parse the responses to be consistent
69
77
  response_data = {}
70
78
  response_data["message"] = response.get("message")
79
+ if response.get("tool_names") and response.get("tool_args_list"):
80
+ response_data["tool_names"] = response.get("tool_names")
81
+ response_data["tool_args_list"] = response.get("tool_args_list")
71
82
  response_data["completion_tokens"] = response.get("completion_tokens")
72
83
  response_data["prompt_tokens"] = response.get("prompt_tokens")
73
84
  response_data["response_duration"] = response.get("response_duration")
not_again_ai/local_llm/ollama/chat_completion.py CHANGED
@@ -13,6 +13,7 @@ def chat_completion(
13
13
  messages: list[dict[str, Any]],
14
14
  model: str,
15
15
  client: Client,
16
+ tools: list[dict[str, Any]] | None = None,
16
17
  max_tokens: int | None = None,
17
18
  context_window: int | None = None,
18
19
  temperature: float = 0.8,
@@ -27,6 +28,8 @@ def chat_completion(
27
28
  messages (list[dict[str, Any]]): A list of messages to send to the model.
28
29
  model (str): The model to use.
29
30
  client (Client): The Ollama client.
31
+ tools (list[dict[str, Any]], optional):A list of tools the model may call.
32
+ Use this to provide a list of functions the model may generate JSON inputs for. Defaults to None.
30
33
  max_tokens (int, optional): The maximum number of tokens to generate. Ollama calls this `num_predict`.
31
34
  context_window (int, optional): The number of tokens to consider as context. Ollama calls this `num_ctx`.
32
35
  temperature (float, optional): The temperature of the model. Increasing the temperature will make the model answer more creatively.
@@ -38,6 +41,9 @@ def chat_completion(
38
41
  dict[str, Any]: A dictionary with the following keys
39
42
  message (str | dict): The content of the generated assistant message.
40
43
  If json_mode is True, this will be a dictionary.
44
+ tool_names (list[str], optional): The names of the tools called by the model.
45
+ If the model does not support tools, a ResponseError is raised.
46
+ tool_args_list (list[dict], optional): The arguments of the tools called by the model.
41
47
  prompt_tokens (int): The number of tokens in the messages sent to the model.
42
48
  completion_tokens (int): The number of tokens used by the model to generate the completion.
43
49
  response_duration (float): The time, in seconds, taken to generate the response by using the model.
@@ -59,6 +65,8 @@ def chat_completion(
59
65
  }
60
66
  if json_mode:
61
67
  all_args["format"] = "json"
68
+ if tools:
69
+ all_args["tools"] = tools
62
70
 
63
71
  try:
64
72
  start_time = time.time()
@@ -78,19 +86,26 @@ def chat_completion(
78
86
  response_data: dict[str, Any] = {}
79
87
 
80
88
  # Handle getting the message returned by the model
81
- message = response["message"].get("content", None) # type: ignore
89
+ message = response["message"].get("content", "")
82
90
  if message and json_mode:
83
91
  with contextlib.suppress(json.JSONDecodeError):
84
92
  message = json.loads(message)
85
- if message:
86
- response_data["message"] = message
93
+ response_data["message"] = message
94
+
95
+ # Try getting tool calls
96
+ if response["message"].get("tool_calls"):
97
+ tool_calls = response["message"]["tool_calls"]
98
+ tool_names = [tool_call["function"]["name"] for tool_call in tool_calls]
99
+ tool_args_list = [tool_call["function"]["arguments"] for tool_call in tool_calls]
100
+ response_data["tool_names"] = tool_names
101
+ response_data["tool_args_list"] = tool_args_list
87
102
 
88
103
  tokenizer = load_tokenizer(model)
89
104
  prompt_tokens = num_tokens_from_messages(messages, tokenizer)
90
105
  response_data["prompt_tokens"] = prompt_tokens
91
106
 
92
107
  # Get the number of tokens generated
93
- response_data["completion_tokens"] = response.get("eval_count", None) # type: ignore
108
+ response_data["completion_tokens"] = response.get("eval_count", None)
94
109
  if response_data["completion_tokens"] is None:
95
110
  response_data["completion_tokens"] = num_tokens_in_string(str(response_data["message"]), tokenizer)
96
111
 
not_again_ai/local_llm/ollama/model_mapping.py CHANGED
@@ -7,6 +7,7 @@ i.e. all phi3 models will start with "phi3".
7
7
  OLLAMA_MODEL_MAPPING = {
8
8
  "phi3": "microsoft/Phi-3-mini-4k-instruct",
9
9
  "llama3:": "nvidia/Llama3-ChatQA-1.5-8B", # Using this version to get around needed to accept an agreement to get access to the tokenizer
10
+ "llama3.1": "unsloth/Meta-Llama-3.1-8B-Instruct",
10
11
  "gemma": "google/gemma-1.1-7b-it", # Requires HF_TOKEN set and accepting the agreement on the HF model page
11
12
  "qwen2": "Qwen/Qwen2-7B-Instruct",
12
13
  "granite-code": "ibm-granite/granite-34b-code-instruct",
{not_again_ai-0.10.3.dist-info → not_again_ai-0.11.0.dist-info}/METADATA RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.1
2
2
  Name: not-again-ai
3
- Version: 0.10.3
3
+ Version: 0.11.0
4
4
  Summary: Designed to once and for all collect all the little things that come up over and over again in AI projects and put them in one place.
5
5
  Home-page: https://github.com/DaveCoDev/not-again-ai
6
6
  License: MIT
@@ -24,15 +24,15 @@ Provides-Extra: viz
24
24
  Requires-Dist: jinja2 (>=3.1,<4.0) ; extra == "local-llm"
25
25
  Requires-Dist: loguru (==0.7.2)
26
26
  Requires-Dist: numpy (>=1.26,<2.0) ; extra == "statistics" or extra == "viz"
27
- Requires-Dist: ollama (>=0.2,<0.3) ; extra == "local-llm"
28
- Requires-Dist: openai (>=1.36,<2.0) ; extra == "llm"
27
+ Requires-Dist: ollama (>=0.3,<0.4) ; extra == "local-llm"
28
+ Requires-Dist: openai (>=1.37,<2.0) ; extra == "llm"
29
29
  Requires-Dist: pandas (>=2.2,<3.0) ; extra == "viz"
30
30
  Requires-Dist: python-liquid (>=1.12,<2.0) ; extra == "llm"
31
31
  Requires-Dist: scikit-learn (>=1.5,<2.0) ; extra == "statistics"
32
32
  Requires-Dist: scipy (>=1.14,<2.0) ; extra == "statistics"
33
33
  Requires-Dist: seaborn (>=0.13,<0.14) ; extra == "viz"
34
34
  Requires-Dist: tiktoken (>=0.7,<0.8) ; extra == "llm"
35
- Requires-Dist: transformers (>=4.42,<5.0) ; extra == "local-llm"
35
+ Requires-Dist: transformers (>=4.43,<5.0) ; extra == "local-llm"
36
36
  Project-URL: Documentation, https://github.com/DaveCoDev/not-again-ai
37
37
  Project-URL: Repository, https://github.com/DaveCoDev/not-again-ai
38
38
  Description-Content-Type: text/markdown
@@ -295,7 +295,7 @@ installed package to indicate that inline type annotations should be checked.
295
295
 
296
296
  ## Typos
297
297
 
298
- Checks for typos using [typos](https://github.com/crate-ci/typos)
298
+ Check for typos using [typos](https://github.com/crate-ci/typos)
299
299
 
300
300
  ```bash
301
301
  (.venv) $ nox -s typos
{not_again_ai-0.10.3.dist-info → not_again_ai-0.11.0.dist-info}/RECORD RENAMED
@@ -4,20 +4,20 @@ not_again_ai/base/file_system.py,sha256=KNQmacO4Q__CQuq2oPzWrg3rQO48n3evglc9bNiP
4
4
  not_again_ai/base/parallel.py,sha256=fcYhKBYBWvob84iKp3O93wvFFdXeidljZsShgBLTNGA,3448
5
5
  not_again_ai/llm/__init__.py,sha256=_wNUL6FDaT369Z8W48FsaC_NkcOZ-ib2MMUvnaLOS-0,451
6
6
  not_again_ai/llm/openai_api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
7
- not_again_ai/llm/openai_api/chat_completion.py,sha256=9p024l0popYW08AbRYwriN_ikkUANYNl0JEFlFXuySA,8915
7
+ not_again_ai/llm/openai_api/chat_completion.py,sha256=PRFi5Sl1K5GOgfWDYygHlmS-Ks1ZE6ETBzinZsz5GCc,8954
8
8
  not_again_ai/llm/openai_api/context_management.py,sha256=BJSG100_qw9MeTCZGztDV5CBXjVOxU4x7gyoRlLxWnI,3561
9
9
  not_again_ai/llm/openai_api/embeddings.py,sha256=4OBnxZicrY6q4dQhuPqMdAnifyjwrsKMTDj-kVre0yc,2500
10
10
  not_again_ai/llm/openai_api/openai_client.py,sha256=6pZw2xw9X-ceV22rhApwFJ2tAKCxi-SxkjxBsTBZ2Nw,2470
11
11
  not_again_ai/llm/openai_api/prompts.py,sha256=7cDfvIKCTYM0t5lK34FLLqYf-SR_cynDXIXw3zWDizA,7094
12
12
  not_again_ai/llm/openai_api/tokens.py,sha256=31neIrY66ejJQ10VB3EWnkN00wuw9vMpCS8tsw2WtFg,4392
13
13
  not_again_ai/local_llm/__init__.py,sha256=BsUn39U3QQaw6yomQHfp_HIPHRIBoMAgjcP3CDADx04,882
14
- not_again_ai/local_llm/chat_completion.py,sha256=m0fe7oGPtl46kvAIU6RhHeIvDu_pAWphrAbFY3K1i1E,3407
14
+ not_again_ai/local_llm/chat_completion.py,sha256=buZQGV2sChaSi5cgiAcOd9gi4lAEzFIGGIqV-1qazAc,4174
15
15
  not_again_ai/local_llm/huggingface/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
16
16
  not_again_ai/local_llm/huggingface/chat_completion.py,sha256=Y6uMbxLG8TaMVi3hJGrMl_G9Y1N_0dld5Kv1iqYnoao,2300
17
17
  not_again_ai/local_llm/huggingface/helpers.py,sha256=YPr8KbQ8Ac_Mn_nBcrFuL3bCl-IuDCdaRvYVCocy8Gk,734
18
18
  not_again_ai/local_llm/ollama/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
19
- not_again_ai/local_llm/ollama/chat_completion.py,sha256=a2V_Od-Q-Oy8av9N8Lq2ltBDsVwTS0H7ywA6kTPkvP0,4142
20
- not_again_ai/local_llm/ollama/model_mapping.py,sha256=EYxwcRKvIrQAZ3iBFUMzs3f_eYevFOTP8A9AFmh0Syo,837
19
+ not_again_ai/local_llm/ollama/chat_completion.py,sha256=C8uU-yq7FL9OLdflZVjbNkEOofmD2A3Hcsd8k-59iS4,5053
20
+ not_again_ai/local_llm/ollama/model_mapping.py,sha256=sJqPg97OO68O0k5MFkTjRLIn9gB7gCRAcyUEjxMfizo,891
21
21
  not_again_ai/local_llm/ollama/ollama_client.py,sha256=dktyw7aKFq4EA3dU7Le5UpfsSq3Oh_POmYSrAI4qLi8,765
22
22
  not_again_ai/local_llm/ollama/service.py,sha256=XczbxISTAp4KHnIkqRZaMsfBohH-TAHrjZ8T9x3cRAY,2900
23
23
  not_again_ai/local_llm/ollama/tokens.py,sha256=k7K7enOSuCJRHteDG0x-CbuivQ2uAtKK3e0Jr7-BUd4,3952
@@ -32,8 +32,8 @@ not_again_ai/viz/distributions.py,sha256=OyWwJaNI6lMRm_iSrhq-CORLNvXfeuLSgDtVo3u
32
32
  not_again_ai/viz/scatterplot.py,sha256=5CUOWeknbBOaZPeX9oPin5sBkRKEwk8qeFH45R-9LlY,2292
33
33
  not_again_ai/viz/time_series.py,sha256=pOGZqXp_2nd6nKo-PUQNCtmMh__69jxQ6bQibTGLwZA,5212
34
34
  not_again_ai/viz/utils.py,sha256=hN7gwxtBt3U6jQni2K8j5m5pCXpaJDoNzGhBBikEU28,238
35
- not_again_ai-0.10.3.dist-info/LICENSE,sha256=btjOgNGpp-ux5xOo1Gx1MddxeWtT9sof3s3Nui29QfA,1071
36
- not_again_ai-0.10.3.dist-info/METADATA,sha256=aXiTuKLLH3tLVi6Lz-5J0_OWn2XwATyJRt975_aPQnA,15518
37
- not_again_ai-0.10.3.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
38
- not_again_ai-0.10.3.dist-info/entry_points.txt,sha256=EMJegugnmJUd-jMUA_qIRMIPAasbei8gP6O4-ER0BxQ,61
39
- not_again_ai-0.10.3.dist-info/RECORD,,
35
+ not_again_ai-0.11.0.dist-info/LICENSE,sha256=btjOgNGpp-ux5xOo1Gx1MddxeWtT9sof3s3Nui29QfA,1071
36
+ not_again_ai-0.11.0.dist-info/METADATA,sha256=S7IbgUSScGSxAeKWmyxk4hpy4qS-eLqpmfDfFEZq4Ok,15517
37
+ not_again_ai-0.11.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
38
+ not_again_ai-0.11.0.dist-info/entry_points.txt,sha256=EMJegugnmJUd-jMUA_qIRMIPAasbei8gP6O4-ER0BxQ,61
39
+ not_again_ai-0.11.0.dist-info/RECORD,,