Author: Wooheon Hong, Jaehyun Lee
Date: 2020.12.21 ~ 2021. 03. 26
This is biomedical entity relation network.
Identify the connection relationship between entitiy in the papers in PubMed.
Find the relationship between entities based on one sentence, consecutive sentences(in this setting, Swanson ABC model can be used) or entire papers.
It can supports relation network for three or more queries.
- python3
- see
requirements.txt
conda create โn name python=3.7
conda activate name
pip install --upgrade pip
pip install -r requirements.txt
We used PubMed thesis data by crwaling
python app.py
test_model.py
provides more arguments than app.py
provides.
python test_model.py
-
load_or_not
- Select whether run model or use previously saved data
-
query
- ex) macrolactin, iridovirus
-
n_papers
- The number of pepers
-
n_sentence(In the app.py, relation unit)
- 0: document
- 1: network consists of one sentence
- 2: network consists of two consecutive sentence. The Swanson ABC model is available.
The amount of entity pairs is large in order 0 > 2 > 1.
-
is_swanson
- Using Swanson ABC model
- n_sentence must be 2
-
type_scispacy
- Select Biomedical Entity Linking model
- Scispacy
-
layout_algo
- Node Coordinate Algorithm
-
laout_dim
- Select Plot Dimension 2, 3
-
radius
- Select hop except path
-
n_nodes
- The number of nodes
-
cutoff
- Select max path(link between query) length
-
n_path
- The number of paths
The data.sqlite
file is the database and stores the username, password, and email address in the Users table.
To protect user passwords, the password is hashed.