def get_content(DOI, refresh=True, *args, **kwds):
""" Helper function to read file content as xml.
Args:
input_doi (str): DOI of article
*args:
**kwds:
Returns:
Content of returned XML file
"""
if not refresh:
db = open_db_connection()
elsevier = db.elsevier
entries = elsevier.find({"doi": DOI})
if len(entries):
if len(entries) > 1:
print(
"More than one entry for given DOI! Only using only first entry."
)
entry = entries[0]
if entry["collected"]:
content = entry["xml"]
return content
content = download(*args, **kwds).text
return content
def search_for_material(material, search):
db = open_db_connection()
if search:
results = db.abstracts.find({"$text": {"$search": search}, "chem_mentions.names": material}, ["year"])
else:
results = db.abstracts.find({"chem_mentions.names": material}, ["year"])
return list(results)
def search_for_topic(search):
db = open_db_connection(db="matstract_db")
if search:
results = db.abstracts.find({"$or": [{"title": {"$regex": ".*{}.*".format(search)}},
{"abstract": {"$regex": ".*{}.*".format(search)}}]}, ["year"])
print(results.count())
return list(results)
else:
return []
def get_keywords(material):
db = open_db_connection(db="tri_abstracts")
print(db.info)
parser = SimpleParser()
material = parser.matgen_parser(material)
print("number of materials is", db.keywords.count())
keywords = db.keywords.find_one({'material': material})
if keywords is not None:
tf = keywords['keywords_tf']
tf_arranged = arrange_keywords(tf)
tfidf = keywords['keywords_tfidf']
tfidf_arranged = arrange_keywords(tfidf)
df = pd.DataFrame()
df['tf'] = tf_arranged
df['tfidf'] = tfidf_arranged
return generate_table(df)
else:
return "No keywords for the specified material"
def check_scopus_collection(year, issn):
"""
Checks the scopus_log collection on MongoDB whether the data for a given year/journal combination
has been collected.
Args:
year: (str) year
issn: (str) issn of journal
Returns:
(bool) True if status of the year/journal pair is "complete"
"""
db = open_db_connection()
log = db.elsevier_log
entry = log.find({"year": year, "issn": issn})[0]
if entry["status"] == "complete":
return True
elif entry["status"] == "incomplete":
return False
else:
raise KeyError("Entry has no status!")
import dash_html_components as html
import dash_core_components as dcc
import pandas as pd
import random
from matstract.utils import open_db_connection, open_es_client
from matstract.extract import parsing
from bson import ObjectId
db = open_db_connection(db="matstract_db")
client = open_es_client()
def random_abstract():
random_document = list(db.abstracts.aggregate([{
"$sample": {
"size": 1
}
}]))[0]
return random_document['abstract']
def sort_results(results, ids):
results_sorted = sorted(results, key=lambda k: ids.index(k['_id']))
return results_sorted
def highlight_material(body, material):
highlighted_phrase = html.Mark(material)
if len(material) > 0 and material in body:
chopped = body.split(material)
newtext = []
import dash_html_components as html
import dash_materialsintelligence as dmi
from flask_caching import Cache
from flask import send_from_directory
from matstract.web.view import annotate_app, similar_app, \
search_app, keyword_app, extract_app, trends_app
from dash.dependencies import Input, Output, State
from matstract.web.callbacks import search_callbacks, annotate_callbacks, \
extract_callbacks, keyword_callbacks, trends_callbacks, similar_callbacks
from matstract.utils import open_db_connection
import os
db = open_db_connection(local=True)
app = dash.Dash()
server = app.server
# To include local css and js files
app.css.config.serve_locally = True
app.scripts.config.serve_locally = True
app.config.suppress_callback_exceptions = True
app.title = "Matstract"
cache = Cache(server, config={"CACHE_TYPE": "simple"})
### CSS settings ###
BACKGROUND = 'rgb(230, 230, 230)'
COLORSCALE = [[0, "rgb(244,236,21)"], [0.3, "rgb(249,210,41)"], [0.4, "rgb(134,191,118)"],
from dash.dependencies import Input, Output, State
import dash_html_components as html
from matstract.web.view import extract_app
from matstract.utils import open_db_connection
db = open_db_connection(local=True, db="matstract_db")
def bind(app):
### Extract App Callbacks ###
@app.callback(Output('extract-highlighted',
'children'), [Input('extract-button', 'n_clicks')],
[State('extract-textarea', 'value')])
def highlight_extracted(n_clicks, text):
if n_clicks is not None:
text, tags = extract_app.extract_ne(text)
spaced_tags = []
for tag in tags:
#spaced_tags += [tag, html.Span()]
span = html.Span(tag)
span.style = {'padding-right': '15px'}
spaced_tags.append(span)
return html.Div(text), html.Br(), html.Div(
html.Label('Extracted Entity tags:')), html.Div(spaced_tags)
@app.callback(
Output('extract-textarea', 'value'),
# Output('similar-textarea', 'value'),
[Input("extract-random", 'n_clicks')])
def get_random(n_clicks):
if n_clicks is not None:
from dash.dependencies import Input, Output, State
from matstract.web.view import similar_app
from matstract.utils import open_db_connection
db = open_db_connection(db="matstract_db", local=True)
def bind(app, cache):
@app.callback(
Output('similar-textarea', 'value'),
# Output('similar-textarea', 'value'),
[Input("similar-random", 'n_clicks')])
def get_random(n_clicks):
if n_clicks is not None:
text = similar_app.random_abstract()
return text
return ""
@cache.memoize(timeout=600)
@app.callback(
Output('similar-table', 'children'),
# [Input('search-box', 'value')])
[Input('similar-button', 'n_clicks')],
[State('similar-textarea', 'value'), State('similar-material-box', 'value')])
def update_table(n_clicks, search, material):
if material is not None:
table = similar_app.generate_table(search, material)
else:
table = similar_app.generate_table(search)
return table
def __init__(self, local=False):
self._db = open_db_connection(access="annotator", local=local)
return kwds_tf
def extract_tfidf(list_of_abstracts, db_l, count=5):
kwds_tfidf = {}
for tt in ['unigrams', 'bigrams', 'trigrams']:
tf = TermFrequency(normalize=True,
first_last_sentence_only=True,
token_type=tt)
tf.fit_tf(list_of_abstracts)
most_common = tf.sorted_frequencies[:20]
idf_scores = [(word, idf_mongo(db_l, word) * score)
for (word, score) in most_common]
top_idf = sorted(idf_scores, key=itemgetter(1), reverse=True)[:count]
kwds_tfidf[tt] = top_idf
return kwds_tfidf
if __name__ == '__main__':
db = open_db_connection()
material = 'GaN'
result = get_search_results(material=material)
abstracts = [doc['abstract'] for doc in result]
# Extract term frequencies
term_frequencies = extract_tf(abstracts, count=5)
# Extract tfidf
tfidf = extract_tfidf(abstracts, db, count=5)
for n_grams in ['unigrams', 'bigrams', 'trigrams']:
print('####', n_grams, '####', sep='\n')
for tf, tf_idf in zip(term_frequencies[n_grams], tfidf[n_grams]):
print(tf, tf_idf)
def contribute(user_creds="matstract/atlas_creds.json",
max_block_size=100,
num_blocks=1):
"""
Gets a incomplete year/journal combination from elsevier_log, queries for the corresponding
dois, and downloads the corresponding xmls for each to the elsevier collection.
Args:
user_creds ((:obj:`str`, optional)): path to contributing user's write-permitted credential file.
max_block_size ((:obj:`int`, optional)): maximum number of articles in block (~1s/article). Defaults to 100.
num_blocks ((:obj:`int`, optional)): maximum number of blocks to run in session. Defaults to 1.
"""
user = json.load(open(user_creds, 'r'))["name"]
db = open_db_connection(user_creds=user_creds)
log = db.elsevier_log
elsevier = db.elsevier
for i in range(num_blocks):
# Verify access at start of each block to detect dropped VPN sessions.
verify_access()
# Get list of all available blocks sorted from largest to smallest.
available_blocks = log.find(
{
"status": "incomplete",
"num_articles": {
"$lt": max_block_size
}
}, ["year", "issn"]).limit(1).sort("num_articles", -1)
# Break if no remaining blocks smaller than max_block_size
if available_blocks.count() == 0:
print(
"No remaining blocks with size <= {}.".format(max_block_size))
break
else:
print("Blocks remaining = {}".format(
min(num_blocks - i, available_blocks.count())))
target = available_blocks[0]
date = datetime.datetime.now().isoformat()
log.update_one(
{
"year": target["year"],
"issn": target["issn"],
"status": "incomplete"
}, {
"$set": {
"status": "in progress",
"updated_by": user,
"updated_on": date
}
})
# Collect scopus for block
print("Collecting entries for Block {}...".format(target["_id"]))
dois = find_articles(year=target["year"],
issn=target["issn"],
get_all=True)
new_entries = collect_entries(dois, user)
# Insert entries into Matstract database
print("Inserting entries into Matstract database...")
for entry in tqdm(new_entries):
if elsevier.find({"doi": entry["doi"]}).count():
elsevier.update_one({"doi": entry["doi"]}, {"$set": entry})
else:
elsevier.insert_one(entry)
# Mark block as completed in log
date = datetime.datetime.now().isoformat()
log.update_one(
{
"year": target["year"],
"issn": target["issn"],
"status": "in progress"
}, {
"$set": {
"status": "complete",
"completed_by": user,
"completed_on": date,
"updated_by": user,
"updated_on": date
}
})
def __init__(self, db_name="matstract_db", local=True):
self._db = open_db_connection(local=local, db=db_name)
