edsl 0.1.36.dev1__tar.gz → 0.1.36.dev3__tar.gz

{edsl-0.1.36.dev1 → edsl-0.1.36.dev3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: edsl
-Version: 0.1.36.dev1
+Version: 0.1.36.dev3
 Summary: Create and analyze LLM-based surveys
 Home-page: https://www.expectedparrot.com/
 License: MIT

edsl-0.1.36.dev3/edsl/__version__.py

@@ -0,0 +1 @@
+__version__ = "0.1.36.dev3"

{edsl-0.1.36.dev1 → edsl-0.1.36.dev3}/edsl/coop/PriceFetcher.py

@@ -16,30 +16,26 @@ class PriceFetcher:
         if self._cached_prices is not None:
             return self._cached_prices
 
+        import os
         import requests
-        import csv
-        from io import StringIO
-
-        sheet_id = "1SAO3Bhntefl0XQHJv27rMxpvu6uzKDWNXFHRa7jrUDs"
-
-        # Construct the URL to fetch the CSV
-        url = f"https://docs.google.com/spreadsheets/d/{sheet_id}/export?format=csv"
+        from edsl import CONFIG
 
         try:
-            # Fetch the CSV data
-            response = requests.get(url)
+            # Fetch the pricing data
+            url = f"{CONFIG.EXPECTED_PARROT_URL}/api/v0/prices"
+            api_key = os.getenv("EXPECTED_PARROT_API_KEY")
+            headers = {}
+            if api_key:
+                headers["Authorization"] = f"Bearer {api_key}"
+            else:
+                headers["Authorization"] = f"Bearer None"
+
+            response = requests.get(url, headers=headers, timeout=20)
             response.raise_for_status()  # Raise an exception for bad responses
 
-            # Parse the CSV data
-            csv_data = StringIO(response.text)
-            reader = csv.reader(csv_data)
-
-            # Convert to list of dictionaries
-            headers = next(reader)
-            data = [dict(zip(headers, row)) for row in reader]
+            # Parse the data
+            data = response.json()
 
-            # self._cached_prices = data
-            # return data
             price_lookup = {}
             for entry in data:
                 service = entry.get("service", None)

{edsl-0.1.36.dev1 → edsl-0.1.36.dev3}/edsl/coop/coop.py

@@ -584,7 +584,10 @@ class Coop:
         )
         self._resolve_server_response(response)
         response_json = response.json()
-        return response_json.get("cost")
+        return {
+            "credits": response_json.get("cost_in_credits"),
+            "usd": response_json.get("cost_in_usd"),
+        }
 
     ################
     # Remote Errors

{edsl-0.1.36.dev1 → edsl-0.1.36.dev3}/edsl/inference_services/registry.py

@@ -11,21 +11,29 @@ from edsl.inference_services.AwsBedrock import AwsBedrockService
 from edsl.inference_services.AzureAI import AzureAIService
 from edsl.inference_services.OllamaService import OllamaService
 from edsl.inference_services.TestService import TestService
-from edsl.inference_services.MistralAIService import MistralAIService
 from edsl.inference_services.TogetherAIService import TogetherAIService
 
-default = InferenceServicesCollection(
-    [
-        OpenAIService,
-        AnthropicService,
-        DeepInfraService,
-        GoogleService,
-        GroqService,
-        AwsBedrockService,
-        AzureAIService,
-        OllamaService,
-        TestService,
-        MistralAIService,
-        TogetherAIService,
-    ]
-)
+try:
+    from edsl.inference_services.MistralAIService import MistralAIService
+
+    mistral_available = True
+except Exception as e:
+    mistral_available = False
+
+services = [
+    OpenAIService,
+    AnthropicService,
+    DeepInfraService,
+    GoogleService,
+    GroqService,
+    AwsBedrockService,
+    AzureAIService,
+    OllamaService,
+    TestService,
+    TogetherAIService,
+]
+
+if mistral_available:
+    services.append(MistralAIService)
+
+default = InferenceServicesCollection(services)

{edsl-0.1.36.dev1 → edsl-0.1.36.dev3}/edsl/jobs/Jobs.py

@@ -219,11 +219,11 @@ class Jobs(Base):
         price_lookup: dict,
         inference_service: str,
         model: str,
-    ):
+    ) -> dict:
         """Estimates the cost of a prompt. Takes piping into account."""
 
         def get_piping_multiplier(prompt: str):
-            """Returns 2 if a prompt includes Jinja brances, and 1 otherwise."""
+            """Returns 2 if a prompt includes Jinja braces, and 1 otherwise."""
 
             if "{{" in prompt and "}}" in prompt:
                 return 2
@@ -231,9 +231,25 @@ class Jobs(Base):
 
         # Look up prices per token
         key = (inference_service, model)
-        relevant_prices = price_lookup[key]
-        output_price_per_token = 1 / float(relevant_prices["output"]["one_usd_buys"])
-        input_price_per_token = 1 / float(relevant_prices["input"]["one_usd_buys"])
+
+        try:
+            relevant_prices = price_lookup[key]
+            output_price_per_token = 1 / float(
+                relevant_prices["output"]["one_usd_buys"]
+            )
+            input_price_per_token = 1 / float(relevant_prices["input"]["one_usd_buys"])
+        except KeyError:
+            # A KeyError is likely to occur if we cannot retrieve prices (the price_lookup dict is empty)
+            # Use a sensible default
+
+            import warnings
+
+            warnings.warn(
+                "Price data could not be retrieved. Using default estimates for input and output token prices. Input: $0.15 / 1M tokens; Output: $0.60 / 1M tokens"
+            )
+
+            output_price_per_token = 0.00000015  # $0.15 / 1M tokens
+            input_price_per_token = 0.00000060  # $0.60 / 1M tokens
 
         # Compute the number of characters (double if the question involves piping)
         user_prompt_chars = len(str(user_prompt)) * get_piping_multiplier(
@@ -258,7 +274,7 @@ class Jobs(Base):
             "cost": cost,
         }
 
-    def estimate_job_cost_from_external_prices(self, price_lookup: dict):
+    def estimate_job_cost_from_external_prices(self, price_lookup: dict) -> dict:
         """
         Estimates the cost of a job according to the following assumptions:
 
@@ -341,7 +357,7 @@ class Jobs(Base):
 
         return output
 
-    def estimate_job_cost(self):
+    def estimate_job_cost(self) -> dict:
         """
         Estimates the cost of a job according to the following assumptions:
 
@@ -357,6 +373,25 @@ class Jobs(Base):
 
         return self.estimate_job_cost_from_external_prices(price_lookup=price_lookup)
 
+    @staticmethod
+    def compute_job_cost(job_results: "Results") -> float:
+        """
+        Computes the cost of a completed job in USD.
+        """
+        total_cost = 0
+        for result in job_results:
+            for key in result.raw_model_response:
+                if key.endswith("_cost"):
+                    result_cost = result.raw_model_response[key]
+
+                    question_name = key.removesuffix("_cost")
+                    cache_used = result.cache_used_dict[question_name]
+
+                    if isinstance(result_cost, (int, float)) and not cache_used:
+                        total_cost += result_cost
+
+        return total_cost
+
     @staticmethod
     def _get_container_class(object):
         from edsl.agents.AgentList import AgentList

{edsl-0.1.36.dev1 → edsl-0.1.36.dev3}/edsl/jobs/runners/JobsRunnerAsyncio.py

@@ -173,19 +173,20 @@ class JobsRunnerAsyncio:
 
         prompt_dictionary = {}
         for answer_key_name in answer_key_names:
-            prompt_dictionary[
-                answer_key_name + "_user_prompt"
-            ] = question_name_to_prompts[answer_key_name]["user_prompt"]
-            prompt_dictionary[
-                answer_key_name + "_system_prompt"
-            ] = question_name_to_prompts[answer_key_name]["system_prompt"]
+            prompt_dictionary[answer_key_name + "_user_prompt"] = (
+                question_name_to_prompts[answer_key_name]["user_prompt"]
+            )
+            prompt_dictionary[answer_key_name + "_system_prompt"] = (
+                question_name_to_prompts[answer_key_name]["system_prompt"]
+            )
 
         raw_model_results_dictionary = {}
+        cache_used_dictionary = {}
         for result in valid_results:
             question_name = result.question_name
-            raw_model_results_dictionary[
-                question_name + "_raw_model_response"
-            ] = result.raw_model_response
+            raw_model_results_dictionary[question_name + "_raw_model_response"] = (
+                result.raw_model_response
+            )
             raw_model_results_dictionary[question_name + "_cost"] = result.cost
             one_use_buys = (
                 "NA"
@@ -195,6 +196,7 @@ class JobsRunnerAsyncio:
                 else 1.0 / result.cost
             )
             raw_model_results_dictionary[question_name + "_one_usd_buys"] = one_use_buys
+            cache_used_dictionary[question_name] = result.cache_used
 
         result = Result(
             agent=interview.agent,
@@ -207,6 +209,7 @@ class JobsRunnerAsyncio:
             survey=interview.survey,
             generated_tokens=generated_tokens_dict,
             comments_dict=comments_dict,
+            cache_used_dict=cache_used_dictionary,
         )
         result.interview_hash = hash(interview)
 

{edsl-0.1.36.dev1 → edsl-0.1.36.dev3}/edsl/results/Result.py

@@ -75,6 +75,7 @@ class Result(Base, UserDict):
         question_to_attributes: Optional[dict] = None,
         generated_tokens: Optional[dict] = None,
         comments_dict: Optional[dict] = None,
+        cache_used_dict: Optional[dict] = None,
     ):
         """Initialize a Result object.
 
@@ -130,6 +131,7 @@ class Result(Base, UserDict):
         self.question_to_attributes = question_to_attributes
         self.generated_tokens = generated_tokens
         self.comments_dict = comments_dict or {}
+        self.cache_used_dict = cache_used_dict or {}
 
         self._combined_dict = None
         self._problem_keys = None
@@ -153,15 +155,15 @@ class Result(Base, UserDict):
             if key in self.question_to_attributes:
                 # You might be tempted to just use the naked key
                 # but this is a bad idea because it pollutes the namespace
-                question_text_dict[
-                    key + "_question_text"
-                ] = self.question_to_attributes[key]["question_text"]
-                question_options_dict[
-                    key + "_question_options"
-                ] = self.question_to_attributes[key]["question_options"]
-                question_type_dict[
-                    key + "_question_type"
-                ] = self.question_to_attributes[key]["question_type"]
+                question_text_dict[key + "_question_text"] = (
+                    self.question_to_attributes[key]["question_text"]
+                )
+                question_options_dict[key + "_question_options"] = (
+                    self.question_to_attributes[key]["question_options"]
+                )
+                question_type_dict[key + "_question_type"] = (
+                    self.question_to_attributes[key]["question_type"]
+                )
 
         return {
             "agent": self.agent.traits

{edsl-0.1.36.dev1 → edsl-0.1.36.dev3}/pyproject.toml

@@ -1,19 +1,28 @@
 [build-system]
 build-backend = "poetry.core.masonry.api"
-requires = ["poetry-core", "setuptools<72.0"]
+requires = [ "poetry-core", "setuptools<72.0",]
 
 [tool.poetry]
-authors = ["Apostolos Filippas <apostolos@expectedparrot.com>", "John Horton <john@expectedparrot.com>", "Robin Horton <robin@expectedparrot.com>"]
-classifiers = ["Topic :: Scientific/Engineering :: Artificial Intelligence", "Topic :: Software Development :: Libraries :: Application Frameworks", "Topic :: Software Development :: Libraries :: Python Modules"]
+authors = [ "Apostolos Filippas <apostolos@expectedparrot.com>", "John Horton <john@expectedparrot.com>", "Robin Horton <robin@expectedparrot.com>",]
+classifiers = [ "Topic :: Scientific/Engineering :: Artificial Intelligence", "Topic :: Software Development :: Libraries :: Application Frameworks", "Topic :: Software Development :: Libraries :: Python Modules",]
 description = "Create and analyze LLM-based surveys"
 documentation = "https://docs.expectedparrot.com"
 homepage = "https://www.expectedparrot.com/"
-include = ["edsl/questions/templates/**/*"]
-keywords = ["LLM", "social science", "surveys", "user research"]
+include = [ "edsl/questions/templates/**/*",]
+keywords = [ "LLM", "social science", "surveys", "user research",]
 license = "MIT"
 name = "edsl"
 readme = "README.md"
-version = "0.1.36.dev1"
+version = "0.1.36.dev3"
+[[tool.poetry.packages]]
+include = "edsl"
+
+[tool.tomlsort]
+all = true
+in_place = true
+spaces_before_inline_comment = 2
+spaces_indent_inline_array = 4
+trailing_comma_inline_array = true
 
 [tool.poetry.dependencies]
 python = ">=3.9.1,<3.13"
@@ -49,9 +58,12 @@ urllib3 = ">=1.25.4,<1.27"
 google-generativeai = "^0.8.2"
 
 [tool.poetry.dependencies.black]
-extras = ["jupyter"]
+extras = [ "jupyter",]
 version = "^24.4.2"
 
+[tool.tomlsort.overrides."tool.poetry.dependencies"]
+table_keys = false
+
 [tool.poetry.group.dev.dependencies]
 coverage = "^7.3.3"
 fastapi = "^0.112.1"
@@ -77,16 +89,3 @@ sphinx-rtd-theme = "^2.0.0"
 toml = "^0.10.2"
 toml-sort = "^0.23.1"
 uvicorn = "^0.30.6"
-
-[[tool.poetry.packages]]
-include = "edsl"
-
-[tool.tomlsort]
-all = true
-in_place = true
-spaces_before_inline_comment = 2
-spaces_indent_inline_array = 4
-trailing_comma_inline_array = true
-
-[tool.tomlsort.overrides."tool.poetry.dependencies"]
-table_keys = false

edsl-0.1.36.dev1/edsl/__version__.py

@@ -1 +0,0 @@
-__version__ = "0.1.36.dev1"