rgwfuncs 0.0.94__py3-none-any.whl → 0.0.96__py3-none-any.whl

rgwfuncs/df_lib.py

@@ -811,7 +811,7 @@ def filter_indian_mobiles(df: pd.DataFrame, mobile_col: str) -> pd.DataFrame:
     """
     Filter and return DataFrame rows containing valid Indian mobile numbers.
 
-    This function processes a DataFrame to extract and retain rows where the specified column matches the typical format for Indian mobile numbers. An Indian mobile number is expected to be a digit-only string starting with 6, 7, 8, or 9, and should have at least 4 distinct digits.
+    This function processes a DataFrame to extract and retain rows where the specified column matches the typical format for Indian mobile numbers. An Indian mobile number is expected to be a 10-digit string starting with 6, 7, 8, or 9, containing only digits, and having at least 4 distinct digits.
 
     Parameters:
     - df (pd.DataFrame): The DataFrame to filter.
@@ -828,6 +828,7 @@ def filter_indian_mobiles(df: pd.DataFrame, mobile_col: str) -> pd.DataFrame:
             df[mobile_col].apply(
                 lambda x: (
                     str(x).isdigit() and
+                    len(str(x)) == 10 and
                     str(x).startswith(('6', '7', '8', '9')) and
                     len(set(str(x))) >= 4
                 )
@@ -2087,11 +2088,12 @@ def sync_dataframe_to_sqlite_database(
 
 def load_fresh_data_or_pull_from_cache(fetch_func: Callable[[], pd.DataFrame], cache_dir: str, file_prefix: str, cache_cutoff_hours: int, dtype: dict = None) -> pd.DataFrame:
     """
-    Retrieve data from a cache if a recent cache file exists, or fetch fresh data, save it to the cache, and return it.
+    Retrieve data from a cache if a recent cache file exists, or fetch fresh data, save it to the cache, remove older cache files, and return it.
 
     This function checks a specified directory for the most recent cache file matching a specified prefix.
     If a recent cache file (within the cutoff time in hours) is found, the data is read from there.
-    Otherwise, it calls the data-fetching function, saves the newly fetched data to a new cache file, and returns it.
+    Otherwise, it calls the data-fetching function, saves the newly fetched data to a new cache file,
+    removes all earlier cache files with the same prefix, and returns the data.
 
     Parameters:
     - fetch_func (typing.Callable[[], pd.DataFrame]):
@@ -2145,4 +2147,12 @@ def load_fresh_data_or_pull_from_cache(fetch_func: Callable[[], pd.DataFrame], c
         cache_filename: str = f"{file_prefix}_{current_time_str}.csv"
         df.to_csv(os.path.join(cache_dir, cache_filename), index=False)
 
+        # Remove all earlier cache files with the same prefix
+        for filename in os.listdir(cache_dir):
+            if filename.startswith(file_prefix) and filename.endswith(".csv") and filename != cache_filename:
+                try:
+                    os.remove(os.path.join(cache_dir, filename))
+                except OSError:
+                    continue
+
     return df

{rgwfuncs-0.0.94.dist-info → rgwfuncs-0.0.96.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rgwfuncs
-Version: 0.0.94
+Version: 0.0.96
 Summary: A functional programming paradigm for mathematical modelling and data science
 Home-page: https://github.com/ryangerardwilson/rgwfunc
 Author: Ryan Gerard Wilson
@@ -1735,7 +1735,7 @@ Processes and saves a DataFrame to an SQLite database, adding a timestamp column
 
 ### 46. `load_fresh_data_or_pull_from_cache`
 
-Retrieves data from a cache if a recent cache file exists, or fetches fresh data, saves it to the cache, and returns it. If the cache is too old or doesn’t exist, it uses a fetching function to get new data, which it caches and returns. An optional `dtype` parameter allows specifying column data types when reading from the cache, preventing issues with mixed-type columns.
+Retrieves data from a cache if a recent cache file exists, or fetches fresh data, saves it to the cache, removes older cache files, and returns it. If the cache is too old or doesn’t exist, it uses a fetching function to get new data, which it caches and returns. When fresh data is fetched and saved, all earlier cache files with the same prefix are deleted to keep the cache directory clean. An optional `dtype` parameter allows specifying column data types when reading from the cache, preventing issues with mixed-type columns.
 
 #### Parameters:
 - **`fetch_func` (typing.Callable[[], pd.DataFrame])**: A callable function that fetches fresh data and returns it as a pandas DataFrame.

{rgwfuncs-0.0.94.dist-info → rgwfuncs-0.0.96.dist-info}/RECORD

@@ -1,12 +1,12 @@
 rgwfuncs/__init__.py,sha256=LSn54Tlyskcb6Wab_wUpPLB6UGMe5LdrB3GU88mDEbU,1712
 rgwfuncs/algebra_lib.py,sha256=rKFITfpWfgdBswnbMUuS41XgndEt-jUVz2ObO_ik7eM,42234
-rgwfuncs/df_lib.py,sha256=SUEjUc8kCELtbQE2luMsBGh18aTWS97Wb5s3RdMcmHc,75750
+rgwfuncs/df_lib.py,sha256=QgEydReadUBbBkyWGaDfB_OjtALGdjKmzp_S9YZlUks,76266
 rgwfuncs/docs_lib.py,sha256=i63NzX-V8cGhikYdtkRGAEe2VcuwpXxDUyTRa9xI7l8,1972
 rgwfuncs/interactive_shell_lib.py,sha256=YN0ZnM5twIsOeDKuOQ9ZGURCvvBX0RZjM4a1vO1C3E8,4281
 rgwfuncs/str_lib.py,sha256=hE0VfP6rhQpczsKyCZvH3G1aMRwngKnkW3NTYCEc0Po,3208
-rgwfuncs-0.0.94.dist-info/licenses/LICENSE,sha256=jLvt20gcUZYB8UOvyBvyKQ1qhYYhD__qP7ZDx2lPFkU,1062
-rgwfuncs-0.0.94.dist-info/METADATA,sha256=K0ehKuNHmsn7IbtRWM8o7_323F8RUnYNtry2QHosFWo,62066
-rgwfuncs-0.0.94.dist-info/WHEEL,sha256=CmyFI0kx5cdEMTLiONQRbGQwjIoR1aIYB7eCAQ4KPJ0,91
-rgwfuncs-0.0.94.dist-info/entry_points.txt,sha256=j-c5IOPIQ0252EaOV6j6STio56sbXl2C4ym_fQ0lXx0,43
-rgwfuncs-0.0.94.dist-info/top_level.txt,sha256=aGuVIzWsKiV1f2gCb6mynx0zx5ma0B1EwPGFKVEMTi4,9
-rgwfuncs-0.0.94.dist-info/RECORD,,
+rgwfuncs-0.0.96.dist-info/licenses/LICENSE,sha256=jLvt20gcUZYB8UOvyBvyKQ1qhYYhD__qP7ZDx2lPFkU,1062
+rgwfuncs-0.0.96.dist-info/METADATA,sha256=s2T3236W7cPHPnCM8dsFHk238-BIfncEBpzjVDToBwc,62223
+rgwfuncs-0.0.96.dist-info/WHEEL,sha256=0CuiUZ_p9E4cD6NyLD6UG80LBXYyiSYZOKDm5lp32xk,91
+rgwfuncs-0.0.96.dist-info/entry_points.txt,sha256=j-c5IOPIQ0252EaOV6j6STio56sbXl2C4ym_fQ0lXx0,43
+rgwfuncs-0.0.96.dist-info/top_level.txt,sha256=aGuVIzWsKiV1f2gCb6mynx0zx5ma0B1EwPGFKVEMTi4,9
+rgwfuncs-0.0.96.dist-info/RECORD,,

{rgwfuncs-0.0.94.dist-info → rgwfuncs-0.0.96.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (78.1.0)
+Generator: setuptools (80.3.1)
 Root-Is-Purelib: true
 Tag: py3-none-any