merge

Author: ChenZiHong-Gavin

.github/sync-config.yml

@@ -5,8 +5,8 @@ sync:
     dest: resources/nltk_data/
   - source: resources/examples/
     dest: resources/examples/
-  - source: resources/images/logo.png
-    dest: resources/images/logo.png
+  - source: assets/logo.png
+    dest: assets/logo.png
   - source: webui/
     dest: webui/
   - source: webui/app.py

README.md

@@ -1,5 +1,5 @@
 <p align="center">
-  <img src="resources/images/logo.png"/>
+  <img src="assets/logo.png"/>
 </p>
 
 <!-- icon -->
@@ -315,7 +315,7 @@ See [analysis](https://deepwiki.com/open-sciencelab/GraphGen) by deepwiki for a
 
 
 ### Workflow
-![workflow](resources/images/flow.png)
+![workflow](assets/flow.png)
 
 
 ## 🍀 Acknowledgements

README_zh.md

@@ -1,5 +1,5 @@
 <p align="center">
-  <img src="resources/images/logo.png"/>
+  <img src="assets/logo.png"/>
 </p>
 
 <!-- icon -->
@@ -313,7 +313,7 @@ GraphGen 首先根据源文本构建细粒度的知识图谱，然后利用期
 
 
 ### 工作流程
-![workflow](resources/images/flow.png)
+![workflow](assets/flow.png)
 
 
 ## 🍀 致谢

resources/images/flow.png assets/flow.pngresources/images/flow.png renamed to assets/flow.png

@@ -9,7 +9,7 @@
 
 from graphgen.bases import BaseLLMWrapper
 from graphgen.common import init_llm
-from graphgen.models import NetworkXStorage
+from graphgen.storage import NetworkXStorage
 from graphgen.utils import create_event_loop
 
 QA_GENERATION_PROMPT = """

resources/images/logo.png assets/logo.pngresources/images/logo.png renamed to assets/logo.png

@@ -10,7 +10,7 @@ nodes:
     dependencies: []
     params:
       input_path:
-        - examples/input_examples/extract_demo.txt
+        - examples/input_examples/jsonl_demo.jsonl
 
   - id: chunk
     op_name: chunk
@@ -39,7 +39,6 @@ nodes:
     dependencies:
       - build_kg
     params:
+      target: kg
       metrics:
-        - kg_structure
-        - kg_accuracy
-        - kg_consistency
+        - structure

baselines/BDS/bds.py

@@ -1,7 +1,7 @@
 global_params:
   working_dir: cache
-  graph_backend: kuzu # graph database backend, support: kuzu, networkx
-  kv_backend: rocksdb # key-value store backend, support: rocksdb, json_kv
+  graph_backend: networkx # graph database backend, support: kuzu, networkx
+  kv_backend: json_kv # key-value store backend, support: rocksdb, json_kv
 
 nodes:
   - id: read_files # id is unique in the pipeline, and can be referenced by other steps
@@ -89,10 +89,11 @@ nodes:
       batch_size: 128
     save_output: true
     params:
+      target: qa
       metrics:
-        - qa_length
-        - qa_mtld
-        - qa_reward_score
-        - qa_uni_score
+        - length
+        - mtld
+        # - reward_score
+        # - uni_score
       mtld_params:
         threshold: 0.7

examples/evaluate/evaluate_kg/kg_evaluation_config.yaml

@@ -0,0 +1,2 @@
+python3 -m graphgen.run \
+--config_file examples/evaluate/evaluate_triple/triple_evaluation_config.yaml

examples/evaluate/evaluate_qa/qa_evaluation_config.yaml

@@ -0,0 +1,46 @@
+global_params:
+  working_dir: cache
+  graph_backend: networkx # graph database backend, support: kuzu, networkx
+  kv_backend: json_kv # key-value store backend, support: rocksdb, json_kv
+
+nodes:
+  - id: read
+    op_name: read
+    type: source
+    dependencies: []
+    params:
+      input_path:
+        - examples/input_examples/jsonl_demo.jsonl
+
+  - id: chunk
+    op_name: chunk
+    type: map_batch
+    dependencies:
+      - read
+    execution_params:
+      replicas: 4
+    params:
+      chunk_size: 20480 # larger chunk size for better context
+      chunk_overlap: 2000
+
+  - id: build_kg
+    op_name: build_kg
+    type: map_batch
+    dependencies:
+      - chunk
+    execution_params:
+      replicas: 1
+      batch_size: 128
+
+  - id: evaluate
+    op_name: evaluate
+    type: aggregate
+    save_output: true
+    dependencies:
+      - build_kg
+    params:
+      target: triple
+      src_namespace: chunk
+      tgt_namespace: build_kg
+      metrics:
+        - accuracy