Fix a few bugs

Author: habedi

Makefile

@@ -144,14 +144,6 @@ clean-all: clean rust-clean ## Clean everything
 check: rust-lint rust-test ## Run all checks (linting and tests)
 	@echo "All checks passed!"
 
-.PHONY: examples
-examples: ## Run SQL examples for Onager extension
-	@echo "Running the examples in the ${EXAMPLES_DIR} directory..."
-	@for sql_file in $(EXAMPLES_DIR)/*.sql; do \
-		echo "Running example: $$sql_file"; \
-		./build/release/duckdb < $$sql_file; \
-		echo "============================================================================"; \
-	done
 
 .PHONY: docs
 docs: ## Generate Onager MkDocs documentation

README.md

@@ -23,7 +23,7 @@ and graph generators.
 Onager is written in Rust and uses [Graphina](https://github.com/habedi/graphina) graph library under the hood.
 
 Compared to [DuckPGQ](https://github.com/cwida/duckpgq-extension), Onager is focused on graph analytics instead of graph querying.
-More specifically, DuckPGQ tries to implement SQL/PGQ (the SQL:2023 standard)
+More specifically, DuckPGQ tries to implement [SQL/PGQ](https://pgql-lang.org/) (the SQL:2023 standard)
 for graph pattern matching and path-finding queries with a property graph model.
 Onager instead provides a collection of ready-to-use graph algorithms (like PageRank, Louvain community detection, Dijkstra's shortest path, etc.)
 as simple table functions.
@@ -32,7 +32,7 @@ Users typically want something like DuckPGQ when they need to query graph patter
 
 ### Features
 
-- Adds over 30 graph algorithms as SQL table functions
+- Adds over 40 graph algorithms as SQL table functions
 - Provides a simple and uniform API
 - Supports both directed and undirected graphs
 - Supports weighted and unweighted edges

ROADMAP.md

@@ -10,7 +10,7 @@ It outlines features to be implemented and their current status.
 
 * [x] PageRank centrality
 * [x] Personalized PageRank
-* [x] Degree centrality (in/out)
+* [x] Degree centrality (in and out)
 * [x] Betweenness centrality
 * [x] Closeness centrality
 * [x] Harmonic centrality
@@ -87,7 +87,7 @@ It outlines features to be implemented and their current status.
 ### 9. Minimum Spanning Tree
 
 * [x] Kruskal's algorithm
-* [x] Prim's algorithm
+* [ ] Prim's algorithm
 
 ### 10. Link Prediction
 

docs/examples/analytics.md

@@ -14,7 +14,7 @@ Use PageRank and degree centrality to identify influential users:
 ```sql
 -- Create social network edges
 create table follows as select * from (values
-  (1, 2), (1, 3), (2, 3), (3, 4), (4, 5), (5, 1), (2, 4), (3, 5)
+  (1::bigint, 2::bigint), (1, 3), (2, 3), (3, 4), (4, 5), (5, 1), (2, 4), (3, 5)
 ) t(follower, followed);
 
 -- Find top influencers by PageRank
@@ -61,8 +61,8 @@ High local clustering may indicate coordinated behavior:
 
 ```sql
 create table transactions as select * from (values
-  (100, 200), (200, 300), (300, 100), -- Triangle (suspicious)
-  (400, 500), (500, 600)               -- Normal chain
+  (100::bigint, 200::bigint), (200, 300), (300, 100), -- Triangle (suspicious)
+  (400, 500), (500, 600)                              -- Normal chain
 ) t(sender, receiver);
 
 -- Find nodes with high clustering (potential fraud rings)
@@ -85,7 +85,7 @@ where triangles > 0;
 ```sql
 -- User interaction graph
 create table interactions as select * from (values
-  (1, 10), (1, 20), (2, 10), (2, 30), (3, 20), (3, 30)
+  (1::bigint, 10::bigint), (1, 20), (2, 10), (2, 30), (3, 20), (3, 30)
 ) t(user_id, item_id);
 
 -- Create edges (user-item bipartite graph)
@@ -95,7 +95,7 @@ create table edges as
 -- Recommend items for user 1 based on their interactions
 -- Using ego graph to explore local neighborhood
 select *
-from onager_sub_k_hop_neighbors(edges, start := 1, k := 2)
+from onager_sub_k_hop((select src, dst from edges), start := 1::bigint, k := 2)
 where node_id > 1000;  -- Filter to items only
 ```
 

docs/examples/basic.md

@@ -1,6 +1,9 @@
-# Basic Usage
+---
+title: Basic Usage
+description: Basic graph analytics workflows using Onager.
+---
 
-This example shows basic graph analytics workflows using the Onager.
+# Basic Usage
 
 ## Setup
 

docs/examples/centrality.md

@@ -1,3 +1,8 @@
+---
+title: Centrality Analysis Example
+description: Finding influential nodes in a network.
+---
+
 # Centrality Analysis Example
 
 ## Finding Influential Nodes in a Network

docs/examples/community.md

@@ -1,3 +1,8 @@
+---
+title: Community Detection Example
+description: Segmenting users into groups.
+---
+
 # Community Detection Example
 
 ## Segmenting Users

docs/examples/links.md

@@ -1,3 +1,8 @@
+---
+title: Link Prediction Example
+description: Friend recommendations using link prediction.
+---
+
 # Link Prediction Example
 
 ## Friend Recommendations
@@ -22,7 +27,7 @@ limit 10;
 ```sql
 with
   aa as (select node1, node2, score as adamic_adar from onager_lnk_adamic_adar((select user1, user2 from current_friends))),
-  jc as (select node1, node2, jaccard from onager_lnk_jaccard((select user1, user2 from current_friends))),
+  jc as (select node1, node2, coefficient as jaccard from onager_lnk_jaccard((select user1, user2 from current_friends))),
   pa as (select node1, node2, score as pref_attach from onager_lnk_pref_attach((select user1, user2 from current_friends)))
 select aa.node1, aa.node2, aa.adamic_adar, jc.jaccard, pa.pref_attach
 from aa

docs/getting-started/quickstart.md

@@ -71,7 +71,7 @@ Calculate distances from a starting node:
 
 ```sql
 select node_id, distance
-from onager_pth_dijkstra((select src, dst from edges), source := 1)
+from onager_pth_dijkstra((select src, dst from edges), source := 1::bigint)
 order by distance;
 ```
 
@@ -80,15 +80,10 @@ order by distance;
 Import your edge data from any source DuckDB supports:
 
 ```sql
--- From a CSV file
-create table my_edges as
-select *
-from read_csv('edges.csv');
-
--- From a Parquet file
-create table my_edges as
-select *
-from read_parquet('edges.parquet');
+-- From an existing table (ensure bigint columns)
+create table my_edges as select * from (values
+  (1::bigint, 2::bigint), (2, 3), (3, 4)
+) t(source_id, target_id);
 
 -- Run PageRank on your data
 select *
@@ -101,9 +96,12 @@ from onager_ctr_pagerank((select source_id as src, target_id as dst
 All functions expect edges as a subquery with two `bigint` columns:
 
 ```sql
+create table your_table as select * from (values (1::bigint, 2::bigint), (2, 3)) t(source_column, target_column);
+
 select *
-from onager_function_name((select source_column as src, target_column as dst
+from onager_ctr_pagerank((select source_column as src, target_column as dst
                            from your_table));
+-- Replace your_table and column names with your actual table and columns
 ```
 
 See the [Input Formats](../reference/input-formats.md) reference for details on node IDs, weighted functions, and directed graphs.

docs/guide/approximation.md

@@ -74,12 +74,13 @@ create table weighted_edges as select * from (values
   (2, 3, 2.0), (3, 4, 1.5), (4, 1, 2.5)
 ) t(src, dst, weight);
 
-select order_idx, node_id
+select "order", node_id
 from onager_apx_tsp((select src, dst, weight from weighted_edges))
-order by order_idx;
+order by "order";
+
 ```
 
-| Column    | Type   | Description              |
-|-----------|--------|--------------------------|
-| order_idx | bigint | Position in the tour     |
-| node_id   | bigint | Node at this position    |
+| Column  | Type   | Description              |
+|---------|--------|--------------------------|
+| order   | bigint | Position in the tour     |
+| node_id | bigint | Node at this position    |