Submission: Arthur Armaing

submissions/Arthur Armaing/README.md

@@ -0,0 +1,12 @@
+EmailToCalendar:
+
+Takes in the last n emails, and scans each email for possible events to add to your calendar. If it detects events, it will connect to your google calendar and add them to it, giving time and location for each event as well as a short description in the title. 
+This is all provided that you have proper credentials to enter your google calendar and a Cloud project.
+
+HENCE, for the sake of running the project, I will be leaving my calendar credentials to this project. THAT SAID, PLEASE DO NOT ABUSE THE CREDENTIALS in ANY WAY to modify, delete, or send calendar events or emails.
+
+The project can be run by running main.py. The default is set to 2 email, but can be changed to more if needed. In the example I will be providing, I will be using 1 email, containing a general body meeting on tuesday at 6:30pm.
+
+Please install the dependencies such as the transformers library, Python, pyTorch, and the provided libraries imported in each file.
+
+The project can be expanded to schedule the task of running the script, and with accuracy of the LLM, both of which I intend to do on my free time.

submissions/Arthur Armaing/calendarTest.py

@@ -0,0 +1,101 @@
+import datetime as dt
+import os.path
+
+from google.auth.transport.requests import Request
+from google.oauth2.credentials import Credentials
+from google_auth_oauthlib.flow import InstalledAppFlow
+from googleapiclient.discovery import build
+from googleapiclient.errors import HttpError
+import requests
+
+SCOPES = ["https://www.googleapis.com/auth/calendar"]
+
+def make_calendar_event(summary, location, start_time):
+    creds = None
+    if os.path.exists("token1.json"):
+        creds = Credentials.from_authorized_user_file("token1.json", SCOPES)
+
+    if not creds or not creds.valid:
+        if creds and creds.expired and creds.refresh_token:
+            creds.refresh(Request())
+        else:
+            flow = InstalledAppFlow.from_client_secrets_file(
+            "credentials.json", SCOPES
+            )
+            creds = flow.run_local_server(port=0)
+
+        # Save the credentials for the next run
+        with open("token1.json", "w") as token:
+            token.write(creds.to_json())
+
+    try:
+        service = build("calendar", "v3", credentials=creds)
+        event = {
+            "summary":summary,
+            "location": location,
+            "description": "",
+            "colorId": 9,
+            "start": {
+                "dateTime": start_time.isoformat(),
+                "timeZone": "America/New_York"
+            },
+            "end": {
+                "dateTime": (start_time + dt.timedelta(hours=1)).isoformat(), # 1 hour long
+                "timeZone": "America/New_York"
+            }
+        }
+        event = service.events().insert(calendarId="primary", body=event).execute()
+        print(f'Event created: {event.get("htmlLink")}')
+
+    except HttpError as error:
+        print(f"An error occurred: {error}")
+
+'''
+if __name__ == "__main__":
+    make_calendar_event( "Python-generated event", "online, duh", dt.datetime.now(dt.timezone(dt.timedelta(hours=-5))))
+'''
+
+
+
+
+
+
+
+
+
+
+
+
+'''
+            "recurrence": {
+                "frequency": "DAILY",
+                "interval": 1
+            },
+            "attendees": {
+                {"email": "arthur.armaing@gmail.com"}
+            }
+'''
+
+
+'''
+        # to list the last 5 events
+        now = dt.datetime.now(dt.timezone.utc).isoformat()
+
+        event_result = service.events().list(
+            calendarId="primary",
+            timeMin=now,
+            maxResults=5,
+            singleEvents=True,
+            orderBy="startTime"
+        ).execute()
+        events = event_result.get("items", [])
+
+        if not events:
+            print("No upcoming events found")
+            return
+
+        for event in events:
+            start = event["start"].get("dateTime") # .get("dateTime", event["start"].get("date"))
+            print(start, " ", event["summary"])
+            print(type(start))
+'''

submissions/Arthur Armaing/classifier.py

@@ -0,0 +1,26 @@
+from transformers import pipeline
+
+def classify_text(text):
+    classifier = pipeline("zero-shot-classification", model="facebook/bart-large-mnli")
+    candidate_labels = ["Event", "Meeting", "Workshop", "Spam", "Advertisement", "Update"]
+    res = classifier(
+        text[:1000],
+        candidate_labels=candidate_labels
+    )
+    print(res['scores'])
+
+    max_score = max(res['scores'])
+    for i in range(len(res['scores'])):
+        if res['scores'][i] == max_score:
+            return candidate_labels[i]
+    return candidate_labels[0]
+
+
+'''
+message = ""
+with open("passage.txt", "r") as file:
+    message = file.read()
+
+print(classify_text(message))
+'''
+

submissions/Arthur Armaing/credentials.json

@@ -0,0 +1 @@
+{"installed":{"client_id":"719470134677-b8e2ujkgbfi8qsd500cmh6fno005vsqp.apps.googleusercontent.com","project_id":"mailtocalendar-471019","auth_uri":"https://accounts.google.com/o/oauth2/auth","token_uri":"https://oauth2.googleapis.com/token","auth_provider_x509_cert_url":"https://www.googleapis.com/oauth2/v1/certs","client_secret":"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","redirect_uris":["http://localhost"]}}

submissions/Arthur Armaing/eviomni_prompt

@@ -0,0 +1,18 @@
+You are a highly skilled knowledge reasoner and extractor. 
+Your task is to carefully read the given question and passages to reason how the passages lead to the answer and extract relevant information that may be used to answer the question.
+
+**Follow these steps:** 
+
+**In the <reason></reason> tag, perform the following steps:**
+*Question Analysis: Analyze the question to understand the specific information they are seeking. Identify the key concepts, entities, and relationships involved.
+*Passage Analysis: For each passage, carefully read and identify sentences or phrases that are useful for answering the given question.
+
+**In the <extract></extract> tag, synthesize useful information from the passages into a coherent narrative. Organize the information logically and concisely.**
+
+**In <answer></answer> tags, give a short answer to the given question, based on the passages, reasoning information, and extracted knowledge. If none of them work, please answer the question based on your knowledge.**
+
+**Question:** 
+{question}
+
+**Passages:** 
+{passages}

submissions/Arthur Armaing/inferenceModel.py

@@ -0,0 +1,74 @@
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import StoppingCriteria, StoppingCriteriaList
+import re
+
+class MultiTokenStoppingCriteria(StoppingCriteria):
+    def __init__(self, stop_ids, device):
+        self.stop_ids = stop_ids
+        self.stop_len = len(stop_ids)
+
+    def __call__(self, input_ids, scores, **kwargs):
+        if len(input_ids[0]) >= self.stop_len:
+            last_tokens = input_ids[0][-self.stop_len:].tolist()
+            return last_tokens == self.stop_ids
+        return False
+
+
+def get_model_inference(event_message):
+    model_name = "HIT-TMG/EviOmni-nq_train-1.5B"
+    model = AutoModelForCausalLM.from_pretrained(
+        model_name,
+        dtype="auto",
+        device_map="auto"
+    )
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
+
+    #formulate prompt and fill in stuff
+    prompt = open("eviomni_prompt", "r").read()
+    passages = event_message
+
+    question = ""
+    with open("prompt.txt", 'r', encoding='utf-8') as file:
+        question = file.read()
+
+    instruction = prompt.format(question=question, passages=passages)
+
+    messages = [
+        {"role": "system", "content": "You are a helpful assistant."},
+        {"role": "user", "content": instruction}
+    ]
+
+    stop_token = "</extract>\n\n"
+    stop_ids = tokenizer.encode(stop_token, add_special_tokens=False)
+
+    stopping_criteria = StoppingCriteriaList([
+        MultiTokenStoppingCriteria(stop_ids, model.device)
+    ])
+
+    text = tokenizer.apply_chat_template(
+        messages,
+        tokenize=False,
+        add_generation_prompt=True
+    )
+    model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
+
+    generated_ids = model.generate(
+        **model_inputs,
+        max_new_tokens=512,
+        stopping_criteria=stopping_criteria
+    )
+    generated_ids = [
+        output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
+    ]
+
+    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
+    match = re.search(r"<extract>(.*?)</extract>", response, re.DOTALL)
+    evidence = match.group(1).strip()
+    return evidence
+
+
+if __name__ == "__main__":
+    with open("passage.txt", "r") as file:
+        msg = file.read()
+    print(get_model_inference(msg))
+

submissions/Arthur Armaing/mailReader.py

@@ -0,0 +1,88 @@
+
+import os.path
+import datetime as dt
+import base64
+
+from google.auth.transport.requests import Request
+from google.oauth2.credentials import Credentials
+from google_auth_oauthlib.flow import InstalledAppFlow
+from googleapiclient.discovery import build
+from googleapiclient.errors import HttpError
+
+# If modifying these scopes, delete the file token.json.
+SCOPES = ["https://www.googleapis.com/auth/gmail.readonly"]
+
+
+def get_last_emails(num_emails):
+  creds = None
+
+  if os.path.exists("token.json"):
+    creds = Credentials.from_authorized_user_file("token.json", SCOPES)
+  # If there are no (valid) credentials available, let the user log in.
+  if not creds or not creds.valid:
+    if creds and creds.expired and creds.refresh_token:
+      creds.refresh(Request())
+    else:
+      flow = InstalledAppFlow.from_client_secrets_file(
+          "credentials.json", SCOPES
+      )
+      creds = flow.run_local_server(port=0)
+
+    with open("token.json", "w") as token:
+      token.write(creds.to_json())
+
+  message_output = []
+  try:
+    now = dt.datetime.now(dt.timezone.utc).isoformat()
+    service = build("gmail", "v1", credentials=creds)
+    results = service.users().messages().list(
+            userId="me",
+            labelIds=["INBOX"],
+            maxResults=num_emails
+        ).execute()
+    messages = results.get("messages", [])
+
+    if not messages:
+        print("No messages found.")
+        return
+
+    for msg in messages:
+        msg_data = service.users().messages().get(userId="me", id=msg["id"]).execute()
+
+        '''headers = msg_data["payload"]["headers"]
+        subject = next((h["value"] for h in headers if h["name"] == "Subject"), "No Subject")
+        sender = next((h["value"] for h in headers if h["name"] == "From"), "Unknown Sender")
+        date = next((h["value"] for h in headers if h["name"] == "Date"), "Unknown Date")'''
+
+        # Try to extract body text
+        body = ""
+        if "data" in msg_data["payload"]["body"]:  # Simple email
+            body = msg_data["payload"]["body"]["data"]
+        elif "parts" in msg_data["payload"]:  # Multipart email
+            for part in msg_data["payload"]["parts"]:
+                if part["mimeType"] == "text/plain" and "data" in part["body"]:
+                    body = part["body"]["data"]
+                    break
+        if body:
+            body = base64.urlsafe_b64decode(body).decode("utf-8", errors="ignore")
+            message_output.append(body[:700])
+        else:
+            body = "[No text body found]"
+
+        '''print(f"From: {sender}")
+        print(f"Subject: {subject}")
+        print(f"Date: {date}")
+        print(f"Body:\n{body[:500]}") 
+        print("-" * 40)'''
+
+  except HttpError as error:
+    print(f"An error occurred: {error}")
+
+  return message_output
+
+
+'''
+if __name__ == "__main__":
+  out = get_last_5_emails()
+  print(out)
+'''

submissions/Arthur Armaing/main.py

@@ -0,0 +1,43 @@
+from mailReader import get_last_emails
+from calendarTest import make_calendar_event
+from classifier import classify_text
+from inferenceModel import get_model_inference, MultiTokenStoppingCriteria
+
+from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import StoppingCriteria, StoppingCriteriaList
+import re
+import datetime as dt
+
+def main():
+    emails = get_last_emails(1)
+
+    for email in emails:
+        if classify_text(email) in ["Meeting", "Event", "Workshop"]:
+            print("Event detected")
+            event_params = get_model_inference(email)
+            event_params = event_params.split(";")
+            # print(event_params)
+            if event_params[1] != "N/A":
+                # Pre-processing of inputs
+                equalsSign = event_params[1].find("=")
+                event_params[1] = event_params[1][equalsSign + 1:].strip()
+                event_params[1] = event_params[1].replace("/", " ").split(" ")
+                #print(event_params[1])
+                event_params[1] = dt.datetime(2025, int(event_params[1][0]), int(event_params[1][1]), int(event_params[1][2]), int(event_params[1][3]))
+
+                equalsSign = event_params[0].find("=")
+                event_params[0] = event_params[0][equalsSign + 1:].strip()
+
+                equalsSign = event_params[2].find("=")
+                event_params[2] = event_params[2][equalsSign + 1:].strip()
+
+                #print(event_params[0])
+                #print(event_params[1])
+                #print(event_params[2])
+
+                make_calendar_event(event_params[0], event_params[2], event_params[1])
+                print("Calendar event created")
+
+
+if __name__ == "__main__":
+    main()

submissions/Arthur Armaing/ml_files/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}