Prep for 0.6.0 release (#511)

* invoke api is not yet live and so llm analysis invoke may fail
But this is also a good practice to let the program continue if it fails

* bumped up version

* PR comments addressed

* Updated tracking keys to prod workspace

* fixed a regression introduced from pb change

.github/workflows/tests.yaml

@@ -90,9 +90,10 @@ jobs:
     strategy:
       matrix:
         warehouse:
-          - test: tests/integration/bigquery/pynative.py
-            site_config: BIGQUERY_SITE_CONFIG
-            python_version: 3.10.13
+          # TODO: Revert once bigquery access issue is fixed
+          # - test: tests/integration/bigquery/pynative.py
+          #   site_config: BIGQUERY_SITE_CONFIG
+          #   python_version: 3.10.13
           - test: tests/integration/snowflake/pynative.py
             site_config: SNOWFLAKE_SITE_CONFIG
             python_version: 3.10.13

src/predictions/profiles_mlcorelib/py_native/attribution_report.py

@@ -819,7 +819,7 @@ def register_dependencies(self, this: WhtMaterial):
                         {selector_sql}
                     {{% endwith %}}
                 {{% endmacro %}}
-                {{% exec %}} {{{{warehouse.CreateReplaceTableAs(this.Name(), selector_sql())}}}} {{% endexec %}}
+                {{% exec %}} {{{{warehouse.CreateReplaceTableAs(this, selector_sql())}}}} {{% endexec %}}
             {{% endmacro %}}
             {{% exec %}} {{{{warehouse.BeginEndBlock(begin_block())}}}} {{% endexec %}}"""
 

src/predictions/profiles_mlcorelib/py_native/id_stitcher/audit.py

@@ -76,7 +76,7 @@ def _set_id_stitcher_model(self, this: WhtMaterial):
             retry_count = 0
             while retry_count < max_retries:
                 selected_model_name = self.reader.get_input(
-                    f"Multiple id_stitcher models found. Please select one {id_stitcher_models.keys()}"
+                    f"Multiple id_stitcher models found. Please select one of {list(id_stitcher_models.keys())}"
                 ).strip()
                 if selected_model_name in id_stitcher_models:
                     break
@@ -135,6 +135,7 @@ def _run(self, this: WhtMaterial):
             self.run_completed = True
             n_visualisations = cluster_report.counter
         except Exception as e:
+            self.logger.warn(f"An error occurred while running the audit: {e}")
             n_visualisations = None
         duration = (datetime.now() - self.start_time).total_seconds()
         analytics.track(

src/predictions/profiles_mlcorelib/py_native/id_stitcher/cluster_report.py

@@ -5,8 +5,7 @@
 from profiles_rudderstack.material import WhtMaterial
 from profiles_rudderstack.logger import Logger
 
-# TODO: Uncomment the following line after adding the Reader class to the profiles_rudderstack package
-# from profiles_rudderstack.reader import Reader
+from profiles_rudderstack.reader import Reader
 import networkx as nx
 from pyvis.network import Network
 import plotly.graph_objects as go
@@ -37,7 +36,7 @@
 class ClusterReport:
     def __init__(
         self,
-        reader,
+        reader: Reader,
         this: WhtMaterial,
         entity: Dict,
         table_report: TableReport,

src/predictions/profiles_mlcorelib/py_native/id_stitcher/consent_manager.py

@@ -8,8 +8,7 @@
     LLM_CONSENT_KEY,
 )
 
-# TODO: Uncomment the following line after adding the Reader class to the profiles_rudderstack package
-# from profiles_rudderstack.reader import Reader
+from profiles_rudderstack.reader import Reader
 
 
 class ConsentManager:
@@ -62,7 +61,7 @@ def save_consent(self, consent: bool) -> None:
         with open(self.preferences_file, "w") as f:
             yaml.dump(existing_preferences, f)
 
-    def prompt_for_consent(self, reader) -> bool:
+    def prompt_for_consent(self, reader: Reader) -> bool:
         retry_count = 0
         while retry_count < 3:
             response = reader.get_input(self.consent_message).lower().strip()

src/predictions/profiles_mlcorelib/py_native/id_stitcher/llm_report.py

@@ -9,8 +9,7 @@
 from .consent_manager import ConsentManager
 from profiles_rudderstack.logger import Logger
 
-# TODO: Uncomment the following line after adding the Reader class to the profiles_rudderstack package
-# from profiles_rudderstack.reader import Reader
+from profiles_rudderstack.reader import Reader
 from enum import Enum
 
 
@@ -55,9 +54,11 @@ def run(self):
             )
             if state == ProgramState.STOP:
                 return
-
-        print("\n\nGenerating id_stitcher analysis:")
-        self._interpret_results_with_llm()
+        try:
+            self._interpret_results_with_llm()
+        except Exception:
+            # If the Analysis of the report fails, we don't want to stop the program.
+            pass
         print("You can now ask questions about the ID Stitcher analysis results.")
         self.run_interactive_session()
 

src/predictions/profiles_mlcorelib/py_native/llm.py

@@ -277,7 +277,7 @@ def query_template_creator(
                     -- attribute values with their corresponding predicted value.
                     LEFT JOIN predicted_attribute b ON {join_condition}
                 {{% endmacro %}}
-                {{% exec %}} {{{{warehouse.CreateReplaceTableAs(this.Name(), selector_sql())}}}} {{% endexec %}}
+                {{% exec %}} {{{{warehouse.CreateReplaceTableAs(this, selector_sql())}}}} {{% endexec %}}
             {{% endmacro %}}
             {{% exec %}} {{{{warehouse.BeginEndBlock(begin_block())}}}} {{% endexec %}}"""
         return query_template

src/predictions/profiles_mlcorelib/utils/constants.py

@@ -8,8 +8,8 @@
 TELEMETRY_CONSENT_SHOWN_KEY = "telemetry_consent_shown"
 LLM_CONSENT_KEY = "llm_consent"
 # TODO: Modify them to a prod settings. These are currently from a testing env
-WRITE_KEY = "2ohOAcLQ7PQaN2s7jNEPFc1elNW"
-DATA_PLANE_URL = "https://rudderstacjpoy.dataplane.rudderstack.com"
+WRITE_KEY = "2oqUj9oDjSSp5dGpiCr5Q5IGZk6"
+DATA_PLANE_URL = "https://rudderstack-dataplane.rudderstack.com"
 
 
 # Propensity model related constants

src/predictions/setup.py

@@ -2,7 +2,7 @@
 from version import version
 
 install_requires = [
-    "profiles_rudderstack>=0.17.0",
+    "profiles_rudderstack>=0.19.0",
     "cachetools>=5.3.2",
     "hyperopt>=0.2.7",
     "joblib>=1.3.2",