Merge pull request #4 from AstraBert/readability

Readability Fixes

.gitignore

@@ -0,0 +1,5 @@
+# cache
+scripts/__pycache__/
+
+# misc
+articles.xml

README.md

@@ -5,8 +5,8 @@ A Telegram bot to retrieve the title, doi, authors and publication date of paper
 You can pull it from GitHub Docker Container registry:
 
 ```bash
-docker pull ghcr.io/astrabert/biomedicalpapersbot:latest
-docker run -p 7860:7860 ghcr.io/astrabert/biomedicalpapersbot:latest
+docker pull ghcr.io/astrabert/biomedicalpapersbot:main
+docker run -p 7860:7860 ghcr.io/astrabert/biomedicalpapersbot:main
 ```
 
 Or you can clone the repository:
@@ -45,7 +45,8 @@ It is a (bio)python-based Gradio bot that searches PubMed and returns the featur
 You can find a snippet code of the functions used to retrieve and parse data from PubMed in [pubmedScraper.py](./scripts/pubmedScraper.py). The workflow is pretty simple:
 
 - `search_pubmed` does the actual webscraping, thanks to the Entrez NCBI module, that remotely connects to online servers and communicate with them: the function returns a list of PubMed IDs
-- `fetch_pubmed_details`, thanks to a faster access to paper metadata and data with the IDs from the previous function, retrieves significant information about papers and outputs it in standard text format
+- `fetch_pubmed_details`, thanks to a faster access to paper metadata and data with the IDs from the previous function, retrieves significant information about papers and outputs it in standard XML format
+- `fetch_xml` takes care of parsing the XML output and extracting titles, authors, dates of publication and DOIs.
 - `respond_to_query` outputs the information of interest in a format that is human-readable and message-sendable
 
 You can also find the basic architecture of the python code that is used for the Gradio bot itself. 

scripts/app.py

@@ -1,5 +1,5 @@
 import gradio as gr
-from .pubmedScraper import respond_to_query
+from pubmedScraper import respond_to_query
 import time
 
 
@@ -22,8 +22,11 @@ def respond(
         gr.Textbox(value="your.email@example.com", label="e-mail address (optional)"),
         gr.Slider(minimum=1, maximum=15, value=5, step=1, label="Maximum number of results"),
     ],
+    title="""<h1 align='center'>BioMedicalPapersBot</h1>
+<h2 align='center'>Scrape PubMed faster, boost your research!🔬</h2>
+<h3 align='center'>[<a href="https://github.com/AstraBert/BioMedicalPapersBot">GitHub⭐</a>] [<a href="https://github.com/sponsors/AstraBert">Funding</a>]</h3>"""
 )
 
 
 if __name__ == "__main__":
-    demo.launch(server_name="0.0.0.0", port=7860)
+    demo.launch(server_name="0.0.0.0", server_port=7860)

scripts/pubmedScraper.py

@@ -1,4 +1,5 @@
 from Bio import Entrez
+import xml.etree.ElementTree as ET
 
 def remove_blankets(ls):
     for i in range(len(ls)):
@@ -19,37 +20,66 @@ def search_pubmed(query, max_results, address):
 
 def fetch_pubmed_details(pubmed_ids, address):
     Entrez.email = address  # Replace with your email
-    handle = Entrez.efetch(db="pubmed", id=pubmed_ids, rettype="medline", retmode="text")
+    handle = Entrez.efetch(db="pubmed", id=pubmed_ids, rettype="medline", retmode="xml")
     records = handle.read()
     handle.close()
-    return records
+    recs = records.decode("utf-8")
+    f = open("articles.xml", "w")
+    f.write(recs)
+    f.close()
+    return "articles.xml"
 
-def respond_to_query(query,address,max_results=10):
+def fetch_xml(xml_file):
+    tree = ET.parse(xml_file)
+    root = tree.getroot()
 
-    # Perform the PubMed search
-    pubmed_ids = search_pubmed(query, max_results,address)
+    articles = {}
+
+    # Iterate over each article and extract title, authors, and DOI
+    for article in root.findall('PubmedArticle'):
+        # Extract the article title
+        title = article.find('.//ArticleTitle').text if article.find('.//ArticleTitle') is not None else "No title"
+
+        # Extract the authors
+        authors = []
+        for author in article.findall('.//Author'):
+            last_name = author.find('LastName').text if author.find('LastName') is not None else ""
+            fore_name = author.find('ForeName').text if author.find('ForeName') is not None else ""
+            authors.append(f"{fore_name} {last_name}".strip())
 
-    # Fetch details for the retrieved PubMed IDs
+        # Extract the DOI
+        doi = None
+        for elocation_id in article.findall('.//ELocationID'):
+            if elocation_id.get('EIdType') == 'doi':
+                doi = elocation_id.text
+                break
+        pub_date = article.find('.//PubDate')
+        if pub_date is not None:
+            year = pub_date.find('Year').text if pub_date.find('Year') is not None else ""
+            month = pub_date.find('Month').text if pub_date.find('Month') is not None else ""
+            day = pub_date.find('Day').text if pub_date.find('Day') is not None else ""
+            publication_date = f"{year}-{month}-{day}".strip("-")
+        else:
+            publication_date = "No publication date"
+        articles.update({doi: {"Title": title, "Authors": authors, "PubDate": publication_date}})
+    return articles
+
+def respond_to_query(query,address,max_results=10):
+    pubmed_ids = search_pubmed(query, max_results,address)
     pubmed_details = fetch_pubmed_details(pubmed_ids,address)
+    articles = fetch_xml(pubmed_details)
+    final_res = ""
+    for doi in articles:
+        auths = [f"- <kbd> {author} </kbd>" for author in articles[doi]["Authors"]] if len(articles[doi]["Authors"]) > 0 else ["- <kbd> No authors listed </kbd>",""]
+        authorrs = '\n'.join(auths)
+        res = f"**Title**: {articles[doi]['Title']}\n**Publication date**: {articles[doi]['PubDate']}\n<details>\n\t<summary><b>Authors</b></summary>\n\n{authorrs}\n\n</details>\n\n**DOI**: [{doi}🔗](https://doi.org/{doi}) \n\n-----------------------\n"
+        final_res+=res
+    return final_res
 
-    pubmed_split=pubmed_details.split("\n")
-    str_container=[]
-    counter=-1
-    for i in pubmed_split:
-        str_container.append({})
-        counter+=1
-        if i.startswith("TI"):
-            str_container[counter].update({"Title (sometimes not complete)": i.replace('TI  - ', '')})
-        if i.startswith("AU  - "):
-            str_container[counter].update({"Author": i.replace('AU  - ', '')})
-        if i.startswith("PHST") and i.endswith("[pubmed]"):
-            str_container[counter].update({"Published on PubMed on": i.replace('PHST- ', '').replace('[pubmed]','')})
-        if i.endswith("[doi]") and i.startswith("AID - "):
-            str_container[counter].update({"doi": f"https://doi.org/{i[6:len(i)-5]}\n"})
-    results=[]
-    for j in str_container:
-        ls=[f"{key}: {j[key]}\n" for key in list(j.keys())]
-        results.append("".join(ls))
-        remove_blankets(results)
-    defstr="".join(results)
-    return defstr
+# if __name__ == "__main__":
+#     pub_ids = search_pubmed("Drosophila evolution over space and time", 5, "astraberte9@gmail.com")
+#     recs = fetch_pubmed_details(pub_ids, "astraberte9@gmail.com")
+#     r = recs.decode("utf-8")
+#     f = open("articles.xml", "w")
+#     f.write(r)
+#     f.close()