fix: assumption that py process runs in project dir

src/predictions/profiles_mlcorelib/py_native/profiles_tutorial/file_generator.py

@@ -11,8 +11,9 @@
 
 
 class FileGenerator:
-    def __init__(self, fast_mode: bool, input_handler: InputHandler):
+    def __init__(self, project_path: str, fast_mode: bool, input_handler: InputHandler):
         self.fast_mode = fast_mode
+        self.project_path = project_path
         self.yaml = YAML()
         self.yaml.preserve_quotes = True
         self.yaml.indent(mapping=2, sequence=4, offset=2)
@@ -90,7 +91,7 @@ def create_pb_project(self, entity_name, id_types, connection_name, id_graph_mod
             "id_types": [{"name": id_type} for id_type in id_types],
         }
 
-        with open(CONFIG_FILE_PATH, "w") as f:
+        with open(os.path.join(self.project_path, CONFIG_FILE_PATH), "w") as f:
             self.yaml.dump(pb_project, f)
 
     def create_inputs_yaml(self, id_mappings):
@@ -109,7 +110,7 @@ def create_inputs_yaml(self, id_mappings):
                 input_entry["app_defaults"]["ids"].append(id_info)
             inputs["inputs"].append(input_entry)
 
-        with open(INPUTS_FILE_PATH, "w") as f:
+        with open(os.path.join(self.project_path, INPUTS_FILE_PATH), "w") as f:
             self.yaml.dump(inputs, f)
 
     def create_profiles_yaml(self, entity_name, tables, model_name):
@@ -132,13 +133,13 @@ def create_profiles_yaml(self, entity_name, tables, model_name):
             ]
         }
 
-        with open(PROFILES_FILE_PATH, "w") as f:
+        with open(os.path.join(self.project_path, PROFILES_FILE_PATH), "w") as f:
             self.yaml.dump(profiles, f)
 
     def validate_shopify_store_id_is_removed(self):
-        with open(CONFIG_FILE_PATH, "r") as file:
+        with open(os.path.join(self.project_path, CONFIG_FILE_PATH), "r") as file:
             pb_project = yaml.safe_load(file)
-        with open(INPUTS_FILE_PATH, "r") as file:
+        with open(os.path.join(self.project_path, INPUTS_FILE_PATH), "r") as file:
             inputs = yaml.safe_load(file)
         if "shopify_store_id" in pb_project["entities"][0][
             "id_types"
@@ -156,10 +157,10 @@ def validate_shopify_store_id_is_removed(self):
 
     def update_bad_anons_filter(self):
         regex_pattern = "(c8bc33a0-7cb7-47f9-b24f-73e077346142|f0ed91a9-e1a9-46a5-9257-d590f45612fe|cbe0ea73-4878-4892-ac82-b9ad42797000|f4690568-e9e7-4182-abc6-6ea2791daba3|b369d6f5-c17a-457c-ab86-5649c1b53883)"
-        with open(CONFIG_FILE_PATH, "r") as file:
+        with open(os.path.join(self.project_path, CONFIG_FILE_PATH), "r") as file:
             pb_project = yaml.safe_load(file)
         for id_type in pb_project["id_types"]:
             if id_type["name"] == "anon_id":
                 id_type["filters"] = [{"type": "exclude", "regex": regex_pattern}]
-        with open(CONFIG_FILE_PATH, "w") as file:
+        with open(os.path.join(self.project_path, CONFIG_FILE_PATH), "w") as file:
             yaml.dump(pb_project, file)

src/predictions/profiles_mlcorelib/py_native/profiles_tutorial/tutorial.py

@@ -56,7 +56,7 @@ def validate(self) -> tuple[bool, str]:
 class TutorialRecipe(PyNativeRecipe):
     def __init__(self, fast_mode: bool) -> None:
         super().__init__()
-        self.profile_builder = ProfileBuilder(self.reader, fast_mode)
+        self.fast_mode = fast_mode
         self.logger = Logger("TutorialRecipe")
 
     def describe(self, this: WhtMaterial) -> Tuple[str, str]:
@@ -69,18 +69,23 @@ def register_dependencies(self, this: WhtMaterial):
         pass
 
     def execute(self, this: WhtMaterial):
+        project_path = this.base_wht_project.project_path()
         if not os.path.exists("sample_data"):
             self.logger.info("unzipping sample data...")
-            unzip_sample_data()
-        self.profile_builder.run()
+            unzip_sample_data(project_path, self.logger)
+        profile_builder = ProfileBuilder(project_path, self.reader, self.fast_mode)
+        profile_builder.run()
 
 
 class ProfileBuilder:
-    def __init__(self, reader, fast_mode: bool):
+    def __init__(self, project_path: str, reader, fast_mode: bool):
+        self.project_path = project_path
         self.config = {}
         self.db_manager = None
         self.input_handler = InputHandler(reader, fast_mode)
-        self.file_generator = FileGenerator(fast_mode, self.input_handler)
+        self.file_generator = FileGenerator(
+            self.project_path, fast_mode, self.input_handler
+        )
         self.fast_mode = fast_mode
 
     def run(self):
@@ -990,18 +995,19 @@ def get_sample_data_path():
     )
 
 
-def unzip_sample_data():
+def unzip_sample_data(project_path: str, logger: Logger):
     zip_file_path = get_sample_data_path()
     # Ensure the zip file exists
     if not os.path.exists(zip_file_path):
         raise Exception(f"Error: {zip_file_path} not found.")
+
     # Unzip the file
     try:
         with zipfile.ZipFile(zip_file_path, "r") as zip_ref:
             for file_info in zip_ref.infolist():
                 if not file_info.filename.startswith("__MACOSX"):
                     zip_ref.extract(file_info, ".")
-        print(f"Successfully extracted {zip_file_path} to the current directory")
+        logger.info(f"Successfully extracted {zip_file_path} to {project_path}")
     except Exception as e:
         raise Exception(f"An error occurred while extracting: {str(e)}")
 

src/predictions/setup.py

@@ -50,4 +50,7 @@
     ],
     install_requires=install_requires,
     include_package_data=True,
+    package_data={
+        "profiles_mlcorelib": ["py_native/profiles_tutorial/sample_data.zip"]
+    },
 )