init

Author: suqingdong

.gitignore

@@ -0,0 +1,4 @@
+*pyc
+.vscode
+__pycache__
+test

MANIFEST.in

@@ -0,0 +1,47 @@
+# NCBI Pubmed Toolkits
+---
+## Requirements
+- Python3
+
+---
+## Installation
+```bash
+pip isntall pypubmed
+```
+
+---
+## Usage
+### `search`
+> search Pubmed with term
+```bash
+```
+
+### `citation`
+> generate citation for given PMID
+```bash
+```
+
+### `translate`
+> translate text with googletrans
+```bash
+```
+
+### `pdf`
+> download pdf if available for given PMID
+```bash
+```
+
+### `config`
+> set api_key etc.
+```bash
+```
+
+### `serve`
+> start a server
+```bash
+```
+
+---
+## Documents
+> https://pypubmed.readthedocs.io/en/latest/
+

README.md

@@ -0,0 +1,48 @@
+# **NCBI EUtils**
+> https://www.ncbi.nlm.nih.gov/books/NBK25497/
+
+## `einfo`
+> https://eutils.ncbi.nlm.nih.gov/entrez/eutils/einfo.fcgi?db=pubmed
+
+## `esearch`
+> https://www.ncbi.nlm.nih.gov/books/NBK25499/#chapter4.ESearch
+>> https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term=ngs[title]&retmax=5&retstart=0
+>> https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term=ngs&retmax=5&retstart=0&field=title
+
+## `efetch`
+> https://www.ncbi.nlm.nih.gov/books/NBK25499/#chapter4.EFetch
+>> https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=17284678,9997&retmode=xml
+
+
+## `elink`
+> https://www.ncbi.nlm.nih.gov/books/NBK25499/#chapter4.ELink
+> - get cited (`{"linkname": "pubmed_pubmed_citedin"}`)
+>> https://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&db=pubmed&id=20210808&cmd=neighbor_score&retmode=json
+> - get pdf url
+>> https://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=pubmed&db=pubmed&cmd=prlinks&id=10210801
+
+
+
+### `retmode` and `rettype`
+> https://www.ncbi.nlm.nih.gov/books/NBK25499/table/chapter4.T._valid_values_of__retmode_and/?report=objectonly
+
+### `API KEY`
+> - E-utils users are allowed 3 requests/second without an API key.
+> - Create an API key to increase your e-utils limit to 10 requests/second.
+> - use with `api_key=API_KEY `
+
+
+---
+# **Journals**
+> https://www.ncbi.nlm.nih.gov/nlmcatalog/journals
+> - https://ftp.ncbi.nlm.nih.gov/pubmed/J_Entrez.gz
+> - https://ftp.ncbi.nlm.nih.gov/pubmed/J_Medline.gz
+
+---
+# **Impact Factor**
+> http://www.greensci.net/
+
+
+- https://dataguide.nlm.nih.gov/eutilities/utilities.html
+- https://www.nlm.nih.gov/bsd/licensee/elements_article_source.html
+- https://www.nlm.nih.gov/bsd/licensee/elements_alphabetical.html

help.md

@@ -0,0 +1,9 @@
+"""====================\n
+NCBI Pubmed ToolKits\n
+====================
+"""
+
+__version__ = '1.0.0'
+
+__author__ = 'suqingdong'
+__author_email__ = '[email protected]'

pypubmed/__init__.py

@@ -0,0 +1,38 @@
+import os
+import sys
+import click
+
+
+from pypubmed.core.search import search, advance_search
+from pypubmed.core.citations import citations
+from pypubmed import __version__, __author__, __author_email__, __doc__
+
+
+PY3 = sys.version_info.major == 3
+if not PY3:
+    reload(sys)
+    sys.setdefaultencoding('utf-8')
+
+
+
+__epilog__ = click.style('''
+contact: {__author__} <{__author_email__}>
+'''.format(**locals()), fg='bright_black')
+
+@click.group(epilog=__epilog__, help=click.style(__doc__, fg='bright_blue', bold=True))
+@click.option('-d', '--debug', help='logging in debug mode', default=False, is_flag=True)
+@click.version_option(version=__version__, prog_name='pypubmed')
+def cli(debug):
+    if debug:
+        os.putenv('DEBUG', True)
+
+
+def main():
+    cli.add_command(search)
+    cli.add_command(advance_search)
+    cli.add_command(citations)
+    cli()
+
+
+if __name__ == '__main__':
+    main()

pypubmed/bin/__init__.py

@@ -0,0 +1,28 @@
+import json
+
+
+class Article(object):
+    def __init__(self, **kwargs):
+        for k, v in kwargs.items():
+            setattr(self, k, v)
+
+    def to_dict(self):
+        return self.__dict__
+
+    def to_json(self, **kwargs):
+        return json.dumps(self.__dict__, ensure_ascii=False, **kwargs)
+
+    @property
+    def fields(self):
+        return list(self.__dict__.keys())
+
+    def __repr__(self):
+        return 'Article[{pmid} - {title}]'.format(**self.__dict__)
+
+
+if __name__ == '__main__':
+    
+    p = Article(pmid=1, issn='1234-5678', title='test')
+    print(p)
+    print(p.to_dict())
+    print(p.to_json())

pypubmed/bin/cli.py

@@ -0,0 +1,120 @@
+"""
+NCBI Citations:
+    - https://pubmed.ncbi.nlm.nih.gov/help/#cite
+    - Is this limited? IP blocking with too much requests?
+    - https://pubmed.ncbi.nlm.nih.gov/{PMID}/citations/
+
+Related Website:
+    - https://www.pmid2cite.com/
+    - https://www.plagiarism.org/article/citation-styles
+    - https://ozzz.org/citation-generator/
+
+Format:
+    - AMA: American Medical Association 
+    - MLA: Modern Language Association
+    - APA: American Psychological Association
+    - NLM: National Library of Medicine
+"""
+import os
+import re
+import json
+
+import click
+
+from pypubmed.core.eutils import Eutils
+from pypubmed.util.web_request import WebRequest
+
+
+def ncbi_citations(pmid, fmt=None):
+    url = 'https://pubmed.ncbi.nlm.nih.gov/{}/citations/'.format(pmid)
+    data = WebRequest.get_response(url).json()
+    if fmt in data:
+        return data[fmt]
+    return data
+
+
+def manual_citations(article, fmt='ama'):
+
+    doi = ' doi:{}'.format(article.doi) if article.doi else ''
+    issue = '({})'.format(article.issue) if article.issue else ''
+
+    # ========================================
+    # # merge the pagination?
+    # # - S1-22; quiz S23-4  =>  S1-S24
+    # # - 548-554            =>  548-54
+    # ========================================
+    # pagination_merge = article.pagination
+
+    if fmt == 'ama':
+        citation = '{authors}. {title} {med_abbr}. {year};{volume}{issue}:{pagination}.{doi}'
+        if len(article.authors) > 5:
+            authors = ', '.join('{} {}'.format(author[0], author[2]) for author in article.authors[:3]) + ' et al'
+        else:
+            authors = ', '.join('{} {}'.format(author[0], author[2]) for author in article.authors)
+    elif fmt == 'mla':
+        first_author = '{}, {}'.format(*article.authors[0][:2])
+        issue = ',{}'.format(article.issue) if article.issue else ''
+        citation = '{first_author} et al. \u201c{title}\u201d {journal} vol. {volume}{issue} ({year}): {pagination}.{doi}'
+    elif fmt == 'apa':
+        doi = ' https://doi.org/{}'.format(article.doi) if article.doi else ''
+        author_all = ', '.join('{}, {}.'.format(each[0], '. '.join(each[2])) for each in article.authors[:-1])
+        last_author = article.authors[-1]
+        if len(last_author) == 1:
+            last_author = last_author[0]
+        else:
+            last_author = '{}, {}.'.format(last_author[0], '. '.join(last_author[2]))
+        citation = '{author_all}, & {last_author} ({year}). {title} {journal}, {volume}{issue}, {pagination}.{doi}'
+    elif fmt == 'nlm':
+        author_list = ', '.join('{} {}'.format(each[0], each[2]) if len(each) >= 3 else ';{}'.format(each[0]) for each in article.authors)
+        doi = doi + '. ' if doi else ' '
+        citation = '{author_list}. {title} {med_abbr}. {pubdate};{volume}{issue}:{pagination}.{doi}PMID:{pmid}'
+        if article.pmc:
+            citation += '; PMCID: {pmc}.'
+
+    citation = citation.format(**dict(article.to_dict(), **locals()))
+
+    return citation
+
+
+__epilog__ = '''
+examples:\n
+    citations 1 2 3\n
+    citations 1 2 3 -f nlm\n
+    citations 1 2 3 -m -f apa\n
+'''
+
+@click.command(epilog=__epilog__,
+               no_args_is_help=True,
+               short_help='generate citations for given pmids',
+               help='Citations Tools')
+@click.option('-m', '--manual', help='cite with manual citations, default with ncbi citations', default=False, is_flag=True)
+@click.option('-f', '--fmt', help='the format of citation', type=click.Choice('ama mla apa nlm'.split()), default='ama')
+@click.option('-k', '--api-key', help='the API_KEY of NCIB', envvar='NCBI_API_KEY')
+@click.argument('pmids', nargs=-1)
+def citations(pmids, manual, fmt, api_key):
+    if not pmids:
+        click.secho('please supply pmids or a file', fg='green')
+        exit(1)
+
+    if os.path.isfile(pmids[0]):
+        pmid_list = open(pmids[0]).read().strip().split()
+    else:
+        pmid_list = pmids
+    
+    if manual:
+        e = Eutils(api_key=api_key)
+        articles = e.efetch(pmid_list)
+        for article in articles:
+            citation = manual_citations(article, fmt=fmt)
+            click.secho(citation, fg='yellow')
+    else:
+        for pmid in pmid_list:
+            citation = ncbi_citations(pmid, fmt=fmt)
+            click.secho(json.dumps(citation, indent=2), fg='green')
+    
+
+
+
+
+if __name__ == '__main__':
+    citations()