📝 (_data_extraction.py): remove unnecessary blank line

♻️ (_data_extraction.py): refactor code to improve readability and maintainability
📝 (safer_python_exec.py): add import statement for black module
♻️ (safer_python_exec.py): refactor code to use black module for code formatting
♻️ (data_from_text.py): refactor code to improve readability and maintainability
♻️ (entities_from_text.py): refactor code to improve readability and maintainability
♻️ (helpers.py): refactor code to improve readability and maintainability
♻️ (datagraph_from_pydantic.py): refactor code to use safer_exec instead of kindofsafe_exec
♻️ (datagraph_from_pydantic.py): refactor code to use format_code function for code formatting

promptedgraphs/_data_extraction.py

@@ -20,7 +20,6 @@
 """
 
 
-
 def create_messages(text, name, labels, custom_system_message=None):
     label_list = list(labels.keys())
 
@@ -127,4 +126,3 @@ def create_functions(is_parent_list, schema, fn_name: str = None):
             },
         }
     ]
-

promptedgraphs/code_execution/safer_python_exec.py

@@ -1,6 +1,8 @@
 import ast
 from logging import getLogger
 
+import black
+
 logger = getLogger(__name__)
 
 
@@ -54,6 +56,10 @@ def is_potentially_unsafe(node):  # sourcery skip: low-code-quality
     return False
 
 
+def format_code(code):
+    return black.format_str(code, mode=black.FileMode())
+
+
 def safer_exec(model_code):
     logger.warning("Executing generated data model code from datamodel_code_generator")
     exec_variables = {}

promptedgraphs/extraction/data_from_text.py

@@ -171,7 +171,9 @@ class UserIntent(BaseModel):
     load_config()
 
     msg = """How can I learn more about your product?"""
-    async for intent in data_from_text(text=msg, output_type=UserIntent, config=Config()):
+    async for intent in data_from_text(
+        text=msg, output_type=UserIntent, config=Config()
+    ):
         print(intent)
 
 

promptedgraphs/extraction/entities_from_text.py

@@ -20,10 +20,11 @@ async def entities_from_text(
 ):
     label_list = sorted(labels.keys())
 
-    LabelEnums = enum.Enum('LabelEnums', {l:i for i,l in enumerate(label_list)})
+    LabelEnums = enum.Enum("LabelEnums", {l: i for i, l in enumerate(label_list)})
 
     class EntityReference(BaseModel):
         """The raw text and label for each entity occurrence"""
+
         text_span: str = Field(
             title="Text Span",
             description="The exact text of referenced entity.",
@@ -47,12 +48,16 @@ class EntityReference(BaseModel):
                 title="Reason",
                 description="A short description of why that label was selected.",
             )
+
     EntityReference.__doc__ += """\n
-    Labels must be one of the following: {label_list}""".format(label_list=label_list + ["==NONE=="])
+    Labels must be one of the following: {label_list}""".format(
+        label_list=label_list + ["==NONE=="]
+    )
 
     EntityReference.__doc__ += """\n
-    Label Definitions:\n""" + "\n".join([f" * {label}: {labels[label]}" for label in label_list])
-
+    Label Definitions:\n""" + "\n".join(
+        [f" * {label}: {labels[label]}" for label in label_list]
+    )
 
     async for er in data_from_text(
         text=text,
@@ -64,7 +69,6 @@ class EntityReference(BaseModel):
         yield er
 
 
-
 async def example():
     load_config()
 

promptedgraphs/helpers.py

@@ -11,7 +11,6 @@ def deep_merge(d1: dict[str, any], d2: dict[str, any]):
             d1[k] += str(v)
 
 
-
 def add_space_before_capital(text):
     return re.sub(r"(?<=[a-z])(?=[A-Z])", " ", text)
 
@@ -35,6 +34,7 @@ def format_fieldinfo(key, v: dict):
         l += f"\n\texamples: {v.get('examples')}"
     return l.rstrip()
 
+
 def remove_nas(data: dict[str, any], nulls: list[str] = None):
     nulls = nulls or ["==NA==", "NA", "N/A", "n/a", "#N/A", "None", "none"]
     for k, v in data.items():

promptedgraphs/sources/datagraph_from_pydantic.py

@@ -12,6 +12,7 @@
 import tqdm
 from pydantic import BaseModel, Field
 
+from promptedgraphs.code_execution.safer_python_exec import format_code
 from promptedgraphs.llms.coding import fix_code
 
 logger = logging.getLogger(__name__)
@@ -34,6 +35,7 @@ def kindofsafe_exec(code, dependencies: dict = None):
     ):  # Loading in code by chunks helps with dependency resolution
         dependencies.update(variables)
         exec(snippet, dependencies, variables)
+    DeprecationWarning("This is a deprecated function: use safer_exec")
     return variables
 
 
@@ -201,10 +203,10 @@ def aggregate_python_files(fdir):
     code = []
     for fname in fdir.glob("*.py"):
         with open(fname) as f:
-            c = black.format_str(f.read(), mode=black.FileMode())
+            c = format_code(f.read())
             code.append(c)
             try:
-                black.format_str("\n\n".join(code), mode=black.FileMode())
+                format_code("\n\n".join(code))
             except Exception:
                 logger.error(f"Static code error in: {fname}")
                 return
@@ -215,7 +217,7 @@ def aggregate_python_files(fdir):
     code = isort.code(code, config=config)
 
     # reformat one more time
-    code = black.format_str(code, mode=black.FileMode())
+    code = format_code(code)
 
     with open(fdir / "_all.py", "w") as f:
         f.write(code)
@@ -248,11 +250,11 @@ async def validate_python_files(fdir):
         history = []
         i = 0
         while i < 4:
-            code = black.format_str(code, mode=black.FileMode())
+            code = format_code(code)
             kindofsafe_exec(code)
             # break
             try:
-                code = black.format_str(code, mode=black.FileMode())
+                code = format_code(code)
                 kindofsafe_exec(code)
                 break
             except Exception as e: