From e0522cbe2104dd09c489a608a2177c9bf1a85795 Mon Sep 17 00:00:00 2001
From: Malte Schokolowski <baw8441@uni-hamburg.de>
Date: Thu, 23 Dec 2021 14:18:23 +0100
Subject: [PATCH] bug fixed for max cit/ref not found while updating
---
.../construct_new_graph/initialize_graph.py | 2 +-
verarbeitung/start_script.py | 12 -----
verarbeitung/update_graph/update_depth.py | 44 +++++++++++--------
verarbeitung/update_graph/update_graph.py | 2 +-
4 files changed, 27 insertions(+), 33 deletions(-)
delete mode 100644 verarbeitung/start_script.py
diff --git a/verarbeitung/construct_new_graph/initialize_graph.py b/verarbeitung/construct_new_graph/initialize_graph.py
index ef17df6..bfc7df2 100644
--- a/verarbeitung/construct_new_graph/initialize_graph.py
+++ b/verarbeitung/construct_new_graph/initialize_graph.py
@@ -75,7 +75,7 @@ def initialize_nodes_list(doi_input_list, search_depth_max, search_height_max, t
for pub_doi in doi_input_list: #iterates over every incoming doi
pub = get_pub(pub_doi, test_var)
if (type(pub) != Publication):
- print(pub)
+ #print(pub)
continue
# checks if publication already exists in nodes
diff --git a/verarbeitung/start_script.py b/verarbeitung/start_script.py
deleted file mode 100644
index 6e5854f..0000000
--- a/verarbeitung/start_script.py
+++ /dev/null
@@ -1,12 +0,0 @@
-import sys
-from pathlib import Path
-from verarbeitung.process_main import Processing
-from verarbeitung.dev_files.print_graph_test import try_known_publications, try_delete_nodes
-
-
-doi_list = []
-doi_list.append('https://pubs.acs.org/doi/10.1021/acs.jcim.9b00249')
-#doi_list.append('https://doi.org/10.1021/acs.jcim.9b00249')
-doi_list.append('https://pubs.acs.org/doi/10.1021/acs.jcim.1c00203')
-doi_list.append('https://doi.org/10.1021/acs.jmedchem.0c01332')
-Processing(doi_list, 2, 2, 'test.json')
diff --git a/verarbeitung/update_graph/update_depth.py b/verarbeitung/update_graph/update_depth.py
index 40fc687..2f0a1d8 100644
--- a/verarbeitung/update_graph/update_depth.py
+++ b/verarbeitung/update_graph/update_depth.py
@@ -18,7 +18,9 @@ sys.path.append("../../")
from verarbeitung.construct_new_graph.add_citations_rec import add_citations
from verarbeitung.construct_new_graph.initialize_graph import complete_inner_edges
+from verarbeitung.get_pub_from_input import get_pub
from .update_edges import back_to_valid_edges
+from input.publication import Publication
def reduce_max_height(max_height):
@@ -62,7 +64,7 @@ def get_old_height_depth():
max_height = max(max_height, pub.group)
return(max_height, max_depth)
-def get_old_max_references(old_depth):
+def get_old_max_references(old_depth, test_var):
'''
:param old_depth: old maximum depth to search for citations
:type old_depth: int
@@ -72,13 +74,14 @@ def get_old_max_references(old_depth):
old_max_references = []
for pub in processed_input_list:
if (abs(pub.group) == old_depth):
- for reference in pub.references:
- for ref_pub in processed_input_list:
- if reference.doi_url == ref_pub.doi_url:
- old_max_references.append(ref_pub)
+ pub = get_pub(pub.doi_url, test_var)
+ if (type(pub) != Publication):
+ #print(pub)
+ continue
+ old_max_references.append(pub)
return(old_max_references)
-def get_old_max_citations(old_height):
+def get_old_max_citations(old_height, test_var):
'''
:param old_height: old maximum height to search for citations
:type old_height: int
@@ -88,10 +91,11 @@ def get_old_max_citations(old_height):
old_max_citations = []
for pub in processed_input_list:
if (abs(pub.group) == old_height):
- for citation in pub.citations:
- for cit_pub in processed_input_list:
- if citation.doi_url == cit_pub.doi_url:
- old_max_citations.append(cit_pub)
+ pub = get_pub(pub.doi_url, test_var)
+ if (type(pub) != Publication):
+ #print(pub)
+ continue
+ old_max_citations.append(pub)
return(old_max_citations)
def update_depth(obj_input_list, input_edges, new_depth, new_height, test_var):
@@ -120,20 +124,22 @@ def update_depth(obj_input_list, input_edges, new_depth, new_height, test_var):
old_height, old_depth = get_old_height_depth()
- # removes publications and links from recursion levels which aren't needed anymore
+ # removes publications and links from recursion levels which aren't needed anymore or adds new ones
if (old_depth > new_depth):
reduce_max_depth(new_depth)
- elif (old_height > new_height):
+ elif (old_depth < new_depth):
+ old_max_references = get_old_max_references(old_depth, test_var)
+ add_citations(processed_input_list, valid_edges, old_max_references, old_depth, new_depth, "Reference", test_var)
+
+ if (old_height > new_height):
reduce_max_height(new_height)
+ elif (old_height < new_height):
+ old_max_citations = get_old_max_citations(old_height, test_var)
+ add_citations(processed_input_list, valid_edges, old_max_citations, old_height, new_height, "Citation", test_var)
+
+
- # adds publications and links for new recursion levels
- elif (old_depth < new_depth):
- old_max_references = get_old_max_references(old_depth)
- add_citations(processed_input_list, valid_edges, old_max_references, old_depth+1, new_depth, "Reference", test_var)
- elif (old_height < new_height):
- old_max_citations = get_old_max_citations(old_height)
- add_citations(processed_input_list, valid_edges, old_max_citations, old_height+1, new_height, "Citation", test_var)
back_to_valid_edges(valid_edges, processed_input_list)
# adds edges between reference group and citation group of known publications
diff --git a/verarbeitung/update_graph/update_graph.py b/verarbeitung/update_graph/update_graph.py
index 7bbb907..7e3ad76 100644
--- a/verarbeitung/update_graph/update_graph.py
+++ b/verarbeitung/update_graph/update_graph.py
@@ -59,7 +59,7 @@ def get_new_input_dois(new_input, test_var):
# retrieves information and adds to new list if successful
pub = get_pub(new_node, test_var)
if (type(pub) != Publication):
- print(pub)
+ #print(pub)
continue
new_input_dois.append(pub.doi_url)
--
GitLab