fix(ragas): remove mutable defaults (#684)

Minor PR to clean the whole codebase from the well-known gotcha of
[mutable default
arguments](https://docs.python-guide.org/writing/gotchas/#mutable-default-arguments)
- which I encourage you guys to be more careful of.

Author: mspronesti

src/ragas/embeddings/base.py

@@ -44,9 +44,7 @@ def set_run_config(self, run_config: RunConfig):
 
 class LangchainEmbeddingsWrapper(BaseRagasEmbeddings):
     def __init__(
-        self,
-        embeddings: Embeddings,
-        run_config: t.Optional[RunConfig] = None
+        self, embeddings: Embeddings, run_config: t.Optional[RunConfig] = None
     ):
         self.embeddings = embeddings
         if run_config is None:

src/ragas/evaluation.py

@@ -15,9 +15,9 @@
     LangchainEmbeddingsWrapper,
     embedding_factory,
 )
-from ragas.llms import llm_factory
 from ragas.exceptions import ExceptionInRunner
 from ragas.executor import Executor
+from ragas.llms import llm_factory
 from ragas.llms.base import BaseRagasLLM, LangchainLLMWrapper
 from ragas.metrics._answer_correctness import AnswerCorrectness
 from ragas.metrics.base import Metric, MetricWithEmbeddings, MetricWithLLM
@@ -42,11 +42,11 @@ def evaluate(
     metrics: list[Metric] | None = None,
     llm: t.Optional[BaseRagasLLM | LangchainLLM] = None,
     embeddings: t.Optional[BaseRagasEmbeddings | LangchainEmbeddings] = None,
-    callbacks: Callbacks = [],
+    callbacks: Callbacks = None,
     is_async: bool = False,
     run_config: t.Optional[RunConfig] = None,
     raise_exceptions: bool = True,
-    column_map: t.Dict[str, str] = {},
+    column_map: t.Optional[t.Dict[str, str]] = None,
 ) -> Result:
     """
     Run the evaluation on the dataset with different metrics
@@ -120,6 +120,9 @@ def evaluate(
     'answer_relevancy': 0.874}
     ```
     """
+    column_map = column_map or {}
+    callbacks = callbacks or []
+
     if dataset is None:
         raise ValueError("Provide dataset!")
 

src/ragas/executor.py

@@ -1,8 +1,8 @@
 from __future__ import annotations
-import sys
 
 import asyncio
 import logging
+import sys
 import threading
 import typing as t
 from dataclasses import dataclass, field
@@ -24,28 +24,31 @@ def runner_exception_hook(args: threading.ExceptHookArgs):
 # set a custom exception hook
 # threading.excepthook = runner_exception_hook
 
+
 def as_completed(loop, coros, max_workers):
     loop_arg_dict = {"loop": loop} if sys.version_info[:2] < (3, 10) else {}
     if max_workers == -1:
         return asyncio.as_completed(coros, **loop_arg_dict)
-    
+
     # loop argument is removed since Python 3.10
     semaphore = asyncio.Semaphore(max_workers, **loop_arg_dict)
+
     async def sema_coro(coro):
         async with semaphore:
             return await coro
-    
+
     sema_coros = [sema_coro(c) for c in coros]
     return asyncio.as_completed(sema_coros, **loop_arg_dict)
 
+
 class Runner(threading.Thread):
     def __init__(
         self,
         jobs: t.List[t.Tuple[t.Coroutine, str]],
         desc: str,
         keep_progress_bar: bool = True,
         raise_exceptions: bool = True,
-        run_config: t.Optional[RunConfig] = None
+        run_config: t.Optional[RunConfig] = None,
     ):
         super().__init__()
         self.jobs = jobs
@@ -59,7 +62,7 @@ def __init__(
         self.futures = as_completed(
             loop=self.loop,
             coros=[coro for coro, _ in self.jobs],
-            max_workers=self.run_config.max_workers
+            max_workers=self.run_config.max_workers,
         )
 
     async def _aresults(self) -> t.List[t.Any]:

src/ragas/llms/base.py

@@ -60,7 +60,7 @@ def generate_text(
         n: int = 1,
         temperature: float = 1e-8,
         stop: t.Optional[t.List[str]] = None,
-        callbacks: Callbacks = [],
+        callbacks: Callbacks = None,
     ) -> LLMResult:
         ...
 
@@ -71,7 +71,7 @@ async def agenerate_text(
         n: int = 1,
         temperature: float = 1e-8,
         stop: t.Optional[t.List[str]] = None,
-        callbacks: Callbacks = [],
+        callbacks: Callbacks = None,
     ) -> LLMResult:
         ...
 
@@ -81,7 +81,7 @@ async def generate(
         n: int = 1,
         temperature: float = 1e-8,
         stop: t.Optional[t.List[str]] = None,
-        callbacks: Callbacks = [],
+        callbacks: Callbacks = None,
         is_async: bool = True,
     ) -> LLMResult:
         """Generate text using the given event loop."""
@@ -119,9 +119,7 @@ class LangchainLLMWrapper(BaseRagasLLM):
     """
 
     def __init__(
-        self,
-        langchain_llm: BaseLanguageModel,
-        run_config: t.Optional[RunConfig] = None
+        self, langchain_llm: BaseLanguageModel, run_config: t.Optional[RunConfig] = None
     ):
         self.langchain_llm = langchain_llm
         if run_config is None:
@@ -134,7 +132,7 @@ def generate_text(
         n: int = 1,
         temperature: float = 1e-8,
         stop: t.Optional[t.List[str]] = None,
-        callbacks: t.Optional[Callbacks] = None,
+        callbacks: Callbacks = None,
     ) -> LLMResult:
         temperature = self.get_temperature(n=n)
         if is_multiple_completion_supported(self.langchain_llm):
@@ -164,7 +162,7 @@ async def agenerate_text(
         n: int = 1,
         temperature: float = 1e-8,
         stop: t.Optional[t.List[str]] = None,
-        callbacks: t.Optional[Callbacks] = None,
+        callbacks: Callbacks = None,
     ) -> LLMResult:
         temperature = self.get_temperature(n=n)
         if is_multiple_completion_supported(self.langchain_llm):
@@ -206,8 +204,7 @@ def set_run_config(self, run_config: RunConfig):
 
 
 def llm_factory(
-    model: str = "gpt-3.5-turbo-16k",
-    run_config: t.Optional[RunConfig] = None
+    model: str = "gpt-3.5-turbo-16k", run_config: t.Optional[RunConfig] = None
 ) -> BaseRagasLLM:
     timeout = None
     if run_config is not None:

src/ragas/metrics/base.py

@@ -60,7 +60,8 @@ def save(self, cache_dir: t.Optional[str] = None) -> None:
             "adapt() is not implemented for {} metric".format(self.name)
         )
 
-    def score(self: t.Self, row: t.Dict, callbacks: Callbacks = []) -> float:
+    def score(self: t.Self, row: t.Dict, callbacks: Callbacks = None) -> float:
+        callbacks = callbacks or []
         rm, group_cm = new_group(
             self.name, inputs=row, callbacks=callbacks, is_async=False
         )
@@ -78,8 +79,9 @@ def score(self: t.Self, row: t.Dict, callbacks: Callbacks = []) -> float:
         return score
 
     async def ascore(
-        self: t.Self, row: t.Dict, callbacks: Callbacks = [], is_async: bool = True
+        self: t.Self, row: t.Dict, callbacks: Callbacks = None, is_async: bool = True
     ) -> float:
+        callbacks = callbacks or []
         rm, group_cm = new_group(
             self.name, inputs=row, callbacks=callbacks, is_async=True
         )

src/ragas/testset/docstore.py

@@ -78,7 +78,7 @@ class Direction(str, Enum):
     PREV = "prev"
     UP = "up"
     DOWN = "down"
-    
+
 
 class Node(Document):
     keyphrases: t.List[str] = Field(default_factory=list, repr=False)
@@ -240,7 +240,7 @@ def add_nodes(self, nodes: t.Sequence[Node], show_progress=True):
                 )
                 result_idx += 1
 
-            if n.keyphrases == []:
+            if not n.keyphrases:
                 nodes_to_extract.update({i: result_idx})
                 executor.submit(
                     self.extractor.extract,
@@ -250,7 +250,7 @@ def add_nodes(self, nodes: t.Sequence[Node], show_progress=True):
                 result_idx += 1
 
         results = executor.results()
-        if results == []:
+        if not results:
             raise ExceptionInRunner()
 
         for i, n in enumerate(nodes):
@@ -336,7 +336,6 @@ def adjustment_factor(wins, alpha):
     def get_similar(
         self, node: Node, threshold: float = 0.7, top_k: int = 3
     ) -> t.Union[t.List[Document], t.List[Node]]:
-        items = []
         doc = node
         if doc.embedding is None:
             raise ValueError("Document has no embedding.")

src/ragas/testset/evolutions.py

@@ -464,7 +464,7 @@ async def _aevolve(
         # find a similar node and generate a question based on both
         merged_node = self.merge_nodes(current_nodes)
         similar_node = self.docstore.get_similar(merged_node, top_k=1)
-        if similar_node == []:
+        if not similar_node:
             # retry
             new_random_nodes = self.docstore.get_random_nodes(k=1)
             current_nodes = CurrentNodes(