Quartz sync: Aug 4, 2024, 10:36 PM

.gitignore

@@ -9,4 +9,6 @@ tsconfig.tsbuildinfo
 private/
 .replit
 replit.nix
-.vscode
+.vscode
+content/excalibrain.md
+content/Scripts/.env

content/References/computerProgram/SynaLinksHybridAGI.md

@@ -0,0 +1,5 @@
+[🇿](zotero://select/library/items/V387NZWL)
+
+
+# SynaLinks/HybridAGI (2024)
+

content/References/computerProgram/neo4jNaLLM.md

@@ -0,0 +1,5 @@
+[🇿](zotero://select/library/items/7XHYWD58)
+
+
+# neo4j/NaLLM (2024)
+

content/References/computerProgram/rahulnykgraph_maker.md

@@ -0,0 +1,5 @@
+[🇿](zotero://select/library/items/MESBET73)
+
+[[Entries/Individuals/ rahulnyk]] 
+# rahulnyk/graph_maker (2024)
+

content/References/computerProgram/rahulnykknowledge_graph_maker.md

@@ -0,0 +1,5 @@
+[🇿](zotero://select/library/items/UYJXAI9D)
+
+[[Entries/Individuals/ rahulnyk]] 
+# rahulnyk/knowledge_graph_maker (2024)
+

content/References/journalArticle/Information extraction pipelines for knowledge graphs.md

@@ -0,0 +1,5 @@
+[🇿](zotero://select/library/items/369FGT9H)
+
+[[Entries/Individuals/Mohamad Yaser Jaradeh]] [[Entries/Individuals/Kuldeep Singh]] [[Entries/Individuals/Markus Stocker]] [[Entries/Individuals/Andreas Both]] [[Entries/Individuals/Sören Auer]] 
+# Information extraction pipelines for knowledge graphs (2023)
+

content/References/journalArticle/TEXTBOOK TO TRIPLES Creating knowledge graph in the form of triples from AI TextBook.md

@@ -0,0 +1,5 @@
+[🇿](zotero://select/library/items/MVFJ5Z2R)
+
+[[Entries/Individuals/Aman Kumar]] [[Entries/Individuals/Swathi Dinakaran]] 
+# TEXTBOOK TO TRIPLES: Creating knowledge graph in the form of triples from AI TextBook
+

content/References/webpage/How to Convert Any Text Into a Graph of Concepts.md

@@ -0,0 +1,5 @@
+[🇿](zotero://select/library/items/GDR8CVEQ)
+
+[[Entries/Individuals/Rahul Nayak]] 
+# How to Convert Any Text Into a Graph of Concepts (2023)
+

content/Scripts/pdf_to_txt.py

@@ -0,0 +1,17 @@
+import sys
+from pypdf import PdfReader
+
+def pdf_to_txt(pdf_path, txt_path):
+    with open(pdf_path, 'rb') as pdf_file, open(txt_path, 'w', encoding='utf-8') as txt_file:
+        pdf_reader = PdfReader(pdf_file)
+        for page in pdf_reader.pages:
+            txt_file.write(page.extract_text())
+
+if len(sys.argv) < 3:
+    print("Usage: python pdf_to_txt.py input_pdf output_txt")
+    sys.exit(1)
+
+pdf_path = sys.argv[1]
+txt_path = sys.argv[2]
+
+pdf_to_txt(pdf_path, txt_path)

content/Entries/prolog_to_folder.py → content/Scripts/prolog_to_folder.py

@@ -1,26 +1,7 @@
+import sys
 import os
-import re
+from utils import parse_prolog_file
 
-def parse_prolog_file(file_path):
-    with open(file_path, 'r') as file:
-        content = file.read()
-
-    arity_1_predicates = {}
-    arity_2_predicates = []
-
-    # Parse arity 1 predicates
-    for match in re.finditer(r'(\w+)\((\w+)\)\.', content):
-        predicate, parameter = match.groups()
-        if predicate not in arity_1_predicates:
-            arity_1_predicates[predicate] = set()
-        arity_1_predicates[predicate].add(parameter)
-
-    # Parse arity 2 predicates
-    for match in re.finditer(r'(\w+)\((\w+),\s*(\w+)\)\.', content):
-        predicate, param1, param2 = match.groups()
-        arity_2_predicates.append((predicate, param1, param2))
-
-    return arity_1_predicates, arity_2_predicates
 def create_folders_and_files(arity_1_predicates, arity_2_predicates, output_dir):
     os.makedirs(output_dir, exist_ok=True)
 
@@ -58,9 +39,13 @@ def create_folders_and_files(arity_1_predicates, arity_2_predicates, output_dir)
     print("Folders and files created successfully!")
 
 def main():
-    input_file = "./test.pl"  # Change this to your Prolog file name
-    output_dir = "output"    # Change this to your desired output directory
-
+    if len(sys.argv) < 3:
+        print("Usage: python prolog_to_folder.py input_file output_dir")
+        sys.exit(1)
+
+    input_file = sys.argv[1]
+    output_dir = sys.argv[2]
+
     arity_1_predicates, arity_2_predicates = parse_prolog_file(input_file)
     create_folders_and_files(arity_1_predicates, arity_2_predicates, output_dir)
 

content/Scripts/prompts.py

@@ -0,0 +1,71 @@
+def correctness_check_prompt(domain_subjects):
+    return f'''You are a domain expert in the field of {domain_subjects}.
+    Check the Prolog code for correctness based on the text. Ensure all relationships are logically sound and perfectly consistent with the text.
+    If you find any inconsistencies, correct them in the Prolog code.
+
+    Please respond with prolog code only.'''
+
+relation_prompt = '''You are an expert at creating Knowledge Graphs in Prolog. 
+Translate sentences in the text into Prolog code using predicates of arity 2.
+Arity 2 predicates define relationships (verbs) between nouns, they are provided below. 
+
+You can ONLY use the following predicates:
+category/2
+challenges/2
+contributes/2
+creates/2
+does/2
+provokes/2
+targets/2
+parent/2
+
+Please respond with prolog code only.
+
+Example:
+Text: "John creates a project. The project targets education."
+Step-by-Step Translation:
+	a.	Entities and Objects: "John", "project", "education"
+	b.	Relationships: "creates", "targets"
+	c.	Arity 2 Predicates:
+creates(john, project).
+targets(project, education).
+	e.	Validation: Ensure all entities and relationships are included and correctly formatted.
+	f.	Output:
+creates(john, project).
+targets(project, education).
+
+Text:'''
+
+categories_prompt = '''You are an expert at creating Knowledge Graphs in Prolog.
+Categorize the list of entities using Prolog predicates of arity 1. Only assign 1 category to 1 entity.
+
+You can ONLY use the following predicates:
+abstract_concept/1
+physical_entity/1
+country/1
+person/1
+organization/1
+event/1
+theoretical_framework/1
+
+Please respond with prolog code only.
+
+Example:
+Entities: "John", "Cuba", "Mount Everest", "education"
+Step-by-Step Translation:
+    a.	Entities: "John", "Cuba", "Mount Everest", "education"
+    b.	Categories: "person", "country", "physical_entity", "abstract_concept"
+    c.	Arity 1 Predicates:
+person(john).
+country(cuba).
+physical_entity(mount_everest).
+abstract_concept(education).
+    d.	Validation: Ensure all entities are categorized and correctly formatted.
+    e.	Output:
+person(john).
+country(cuba).
+physical_entity(mount_everest).
+abstract_concept(education).
+
+Entities:
+'''

content/Scripts/text_to_prolog.py

@@ -0,0 +1,29 @@
+import sys
+from utils import parse_prolog_predicates, call_gpt_api, relation_correctness_check, file_to_chunks, text_to_relations, prolog_predicates_to_entities, entities_to_categorized_entities
+from prompts import relation_prompt, categories_prompt, correctness_check_prompt
+
+if len(sys.argv) < 3:
+    print("Usage: python text_to_prolog.py input_file domain_subjects")
+    sys.exit(1)
+
+input_file = sys.argv[1] # e.g. 'nieo.txt'
+domain_subjects = sys.argv[2] # e.g. 'NIEO, international relations, economics'
+name = input_file.split('.')[0]
+
+original_text_chunk_size = 2000
+original_text_chunks = file_to_chunks(input_file, original_text_chunk_size)
+
+relation_output_file = f'{name}_relations.pl'
+entities_output_file = f'{name}_entities.pl'
+categories_output_file = f'{name}_categories.pl'
+
+output_relations = text_to_relations(original_text_chunks, relation_output_file, relation_prompt, correctness_check_prompt(domain_subjects))
+
+entity_predicates = parse_prolog_predicates(output_relations)
+
+prolog_predicates_to_entities(entity_predicates, entities_output_file)
+
+entities_chunk_size = 2000
+entities_chunks = file_to_chunks(entities_output_file, entities_chunk_size)
+
+entities_to_categorized_entities(entities_chunks, categories_output_file, categories_prompt)