Merge pull request #23 from mingjerli/feat/add-repr-methods

feat: Add __repr__ methods to QueryUnit, QueryUnitGraph, and Pipeline

Author: mingjerli

README.md

@@ -80,6 +80,37 @@ for table in source_tables:
     print(f"  {table}")
 ```
 
+**Output:**
+```
+ColumnLineageGraph(
+  query_units=[cte:monthly_sales, main]
+  nodes=[
+  monthly_sales.month (layer=cte, type=expression)
+  monthly_sales.total_amount (layer=cte, type=aggregate)
+  monthly_sales.user_id (layer=cte, type=direct_column)
+  orders.amount (layer=input, type=base_column)
+  orders.order_date (layer=input, type=base_column)
+  orders.user_id (layer=input, type=base_column)
+  output.month (layer=output, type=direct_column)
+  output.name (layer=output, type=direct_column)
+  output.total_amount (layer=output, type=direct_column)
+  users.name (layer=input, type=base_column)
+  ]
+  edges=[
+  monthly_sales.month -> output.month (direct_column)
+  monthly_sales.total_amount -> output.total_amount (direct_column)
+  orders.amount -> monthly_sales.total_amount (aggregate)
+  orders.order_date -> monthly_sales.month (expression)
+  orders.user_id -> monthly_sales.user_id (direct_column)
+  users.name -> output.name (direct_column)
+  ]
+)
+------------------------------------------------------------
+2 source tables:
+  users
+  orders
+```
+
 ### Multi-Query Pipeline Lineage
 
 ```python
@@ -135,6 +166,24 @@ for impact in impacts:
     print(f"  {impact}")
 ```
 
+**Output:**
+```
+Pipeline with 3 queries
+------------------------------------------------------------
+Execution order (5 tables):
+  1. source_events
+  2. users
+  3. raw_events
+  4. daily_active_users
+  5. user_summary
+------------------------------------------------------------
+Backward lineage for user_summary.event_count (1 sources):
+  ColumnNode('daily_active_users:raw_events.*')
+------------------------------------------------------------
+Forward lineage for source_events.event_timestamp (1 impacts):
+  ColumnNode('user_summary:user_summary.activity_date')
+```
+
 ### Metadata from SQL Comments
 
 ```python
@@ -176,6 +225,16 @@ for col in pipeline.columns.values():
             print(f"  PII: {col.sql_metadata.pii}")
 ```
 
+**Output:**
+```
+Total columns: 6
+------------------------------------------------------------
+PII columns (1):
+  select:select.email
+    Owner: data-team
+------------------------------------------------------------
+```
+
 ### Metadata Management and Export
 
 ```python
@@ -229,6 +288,25 @@ GraphVizExporter.export_to_file(pipeline, "lineage.dot")
 print("✓ Exported to lineage.json, columns.csv, lineage.dot")
 ```
 
+**Output:**
+```
+📊 Propagating metadata for 8 columns...
+✅ Done! Propagated metadata for 8 columns
+Found 3 PII columns:
+  ColumnNode('raw.orders:raw.orders.user_email')
+    Owner: data-team
+    Tags: contact, sensitive
+  ColumnNode('analytics.revenue:analytics.revenue.user_email')
+    Owner: data-team
+    Tags: contact, sensitive
+  ColumnNode('analytics.revenue:raw.orders.user_email')
+    Owner: data-team
+    Tags: contact, sensitive
+------------------------------------------------------------
+Exporting to multiple formats...
+✓ Exported to lineage.json, columns.csv, lineage.dot
+```
+
 ### LLM-Powered Description Generation
 
 <!-- skip-test -->
@@ -255,7 +333,7 @@ pipeline = Pipeline(queries, dialect="bigquery")
 
 # Configure LLM (Ollama - free, local), or replace to any LangChain Chat models.
 llm = ChatOllama(model="qwen3-coder:30b", temperature=0.3)
-pipeline.column_graph.llm = llm
+pipeline.llm = llm
 
 # Generate descriptions for all columns
 print(f"Generating descriptions for {len(pipeline.columns)} columns...")
@@ -273,6 +351,31 @@ for col in columns_with_descriptions:
     print(f"    {col.description}")
 ```
 
+**Output:**
+```
+Generating descriptions for 12 columns...
+📊 Generating descriptions for 8 columns...
+✅ Done! Generated 8 descriptions
+------------------------------------------------------------
+Generated descriptions for 8 columns:
+  raw.orders:raw.orders.order_id:
+    Unique identifier for each customer order placed in the system per order record.
+  raw.orders:raw.orders.user_email:
+    User email addresses from the orders table, one per order record.
+  raw.orders:raw.orders.amount:
+    Order total amount in USD per customer.
+  raw.orders:raw.orders.order_date:
+    Order date when customers placed their purchases per day.
+  analytics.revenue:analytics.revenue.user_email:
+    User email addresses from order records, one per order entry.
+  analytics.revenue:raw.orders.user_email:
+    User email addresses from order records, one per order entry.
+  analytics.revenue:analytics.revenue.total_revenue:
+    Total revenue aggregated per customer from order amounts in USD.
+  analytics.revenue:raw.orders.amount:
+    Order total amount in USD per customer from raw orders table.
+```
+
 ## Architecture
 
 ### Conceptual Structure

src/clgraph/models.py

@@ -96,6 +96,9 @@ def __eq__(self, other):
             return False
         return self.unit_id == other.unit_id
 
+    def __repr__(self):
+        return f"{self.unit_type.value}:{self.unit_id}"
+
     def is_leaf(self) -> bool:
         """Check if this is a leaf unit (only depends on base tables)"""
         return len(self.depends_on_units) == 0
@@ -147,6 +150,13 @@ def get_unit_by_name(self, name: str) -> Optional[QueryUnit]:
                 return unit
         return None
 
+    def __repr__(self):
+        """Show topologically sorted query units"""
+        sorted_units = self.get_topological_order()
+        unit_reprs = [repr(unit) for unit in sorted_units]
+        units_str = ", ".join(unit_reprs)
+        return f"QueryUnitGraph([{units_str}])"
+
 
 # ============================================================================
 # Column Lineage Models

src/clgraph/pipeline.py

@@ -1341,6 +1341,22 @@ def _remap_query_ids(self):
             table.read_by = [self.query_mapping.get(qid, qid) for qid in table.read_by]
             table.modified_by = [self.query_mapping.get(qid, qid) for qid in table.modified_by]
 
+    def __repr__(self):
+        """Show topologically sorted SQL statements"""
+        sorted_query_ids = self.table_graph.topological_sort()
+        query_strs = []
+
+        for query_id in sorted_query_ids:
+            query = self.table_graph.queries[query_id]
+            # Truncate SQL to first 60 chars for readability
+            sql_preview = query.sql.strip().replace("\n", " ")
+            if len(sql_preview) > 60:
+                sql_preview = sql_preview[:57] + "..."
+            query_strs.append(f"{query_id}: {sql_preview}")
+
+        queries_display = "\n  ".join(query_strs)
+        return f"Pipeline(\n  {queries_display}\n)"
+
     def split(self, sinks: List) -> List["Pipeline"]:
         """
         Split pipeline into non-overlapping subpipelines based on target tables.