Write some tools to do stuff with papers

app.py

@@ -1,22 +1,112 @@
-from smolagents import CodeAgent,DuckDuckGoSearchTool, HfApiModel,load_tool,tool
+from smolagents import CodeAgent,DuckDuckGoSearchTool, HfApiModel,load_tool,tool, Tool
 import datetime
 import requests
 import pytz
 import yaml
+import json
 from tools.final_answer import FinalAnswerTool
 
 from Gradio_UI import GradioUI
 
-# Below is an example of a tool that does nothing. Amaze us with your creativity !
-@tool
-def my_custom_tool(arg1:str, arg2:int)-> str: #it's import to specify the return type
-    #Keep this format for the description / args / args description but feel free to modify the tool
-    """A tool that does nothing yet 
-    Args:
-        arg1: the first argument
-        arg2: the second argument
+from smolagents import Tool
+from epmc_xml import fetch
+import requests
+import json
+
+class EuropePMCCitationsTool(Tool):
+    name = "europepmc_citation_downloader"
+    description = """
+    This tool queries the Europe PMC API to retrieve citations for a given PMCID.
+    It returns a list of cited PMCID strings from the citation list.
+    """
+    inputs = {
+        "pmcid": {
+            "type": "string",
+            "description": "the PMCID to query (e.g., 'PMC7779037')",
+        }
+    }
+    output_type = "array"
+
+    def forward(self, pmcid: str):
+        url = f"https://www.ebi.ac.uk/europepmc/webservices/rest/PMC/{pmcid}/citations?page=1&pageSize=1000&format=json"
+        
+        headers = {
+            'Content-Type': 'application/json',
+        }
+        
+        response = requests.get(url, headers=headers)
+        if response.status_code != 200:
+            raise Exception(f"Error querying Europe PMC API: {response.status_code}")
+        
+        data = json.loads(response.text)
+        citations = data.get("citationList", {}).get("citation", [])
+        
+        # Extract PMCID strings from citations
+        pmcid_list = [str(citation.get("id", "")) for citation in citations]
+        
+        return pmcid_list
+
+
+class EuropePMCArticleTextTool(Tool):
+    name = "europepmc_fulltext_downloader"
+    description = """
+    This is a tool that fetches the full text of an article from Europe PMC.
+    It takes a PMCID (format PMC\d+) as input and returns the full text of the article as a string.
+    Articles can only be fetched if they are open access; returns emptty string if article is not open access.
     """
-    return "What magic will you build ?"
+    inputs = {
+        "pmcid": {
+            "type": "string",
+            "description": "the PMCID of the article to fetch (e.g., 'PMC7779037')",
+        }
+    }
+    output_type = "string"
+
+    def forward(self, pmcid: str):
+        # Fetch the article using epmc_xml
+        try:
+            article = fetch.article(pmcid)
+        
+            # Extract the full text from the article object
+            full_text = article.get_body()
+        except:
+            full_text = ""#f"The article {pmcid} does not appear to be open access."
+        
+        
+        # Return the full text as a string
+        return full_text
+
+
+class PMID2PMCIDConverter(Tool):
+    name = "pmid_2_pmcid_converter"
+    description = """
+    This tool converts the pmids (all numbers) to pmcids (format PMC\d+).
+    Use this to convert pmids reported by the citation tool to pmcids needed by the fulltext downloader
+    """
+    inputs = {
+        "pmid": {
+            "type": "string",
+            "description": "the PMID to query (e.g., '36033386')",
+        }
+    }
+    output_type = "string"
+
+    def forward(self, pmid):
+        url = f"https://www.ncbi.nlm.nih.gov/pmc/utils/idconv/v1.0?ids={pmid}&versions=no&format=json"
+
+        headers = {
+            'Content-Type': 'application/json',
+        }
+        
+        response = requests.get(url, headers=headers)
+        if response.status_code != 200:
+            raise Exception(f"Error querying NCBI API: {response.status_code}")
+        
+        data = json.loads(response.text)    
+        pmcid = data.get("records", {})[0].get("pmcid", [])
+        
+        return pmcid
+
 
 @tool
 def get_current_time_in_timezone(timezone: str) -> str:
@@ -36,6 +126,10 @@ def get_current_time_in_timezone(timezone: str) -> str:
 
 final_answer = FinalAnswerTool()
 
+citationtool = EuropePMCCitationsTool()
+fulltexttool = EuropePMCArticleTextTool()
+convert_id_tool = PMID2PMCIDConverter()
+
 # If the agent does not answer, the model is overloaded, please use another model or the following Hugging Face Endpoint that also contains qwen2.5 coder:
 # model_id='https://pflgm2locj2t89co.us-east-1.aws.endpoints.huggingface.cloud' 
 
@@ -55,7 +149,7 @@ with open("prompts.yaml", 'r') as stream:
     
 agent = CodeAgent(
     model=model,
-    tools=[final_answer], ## add your tools here (don't remove final answer)
+    tools=[final_answer, citationtool, fulltexttool, convert_id_tool], ## add your tools here (don't remove final answer)
     max_steps=6,
     verbosity_level=1,
     grammar=None,