new tip

index.md

@@ -272,6 +272,10 @@ This file gets generated by [this script](index.py).
 
 - [There is a module for everything in python!](notes/20220904164714.md)
 
+## Marvinai
+
+- [Using marvin ai to extract geo locations](notes/20240628165703.md)
+
 ## Matchcase
 
 - [Use match/case over if/elisf/else](notes/20240614124008.md)

notes/20240628165703.md

@@ -0,0 +1,66 @@
+# Using marvin ai to extract geo locations
+
+I just used Marvin AI to get geo locations based on a fuzzy location string 💡🎉
+
+I like the combination of using a Pydantic 🐍 base model + typing + a prompt in the docstring of a function, feels really clean. 😍  📈
+
+```
+import marvin
+from pydantic import BaseModel, Field
+
+
+class Location(BaseModel):
+    fuzzy_input_name: str = Field(..., description="The original input string")
+    location_string_extracted: str = Field(..., description="The extracted location string")
+    latitude: float = Field(..., description="Latitude of the location")
+    longitude: float = Field(..., description="Longitude of the location")
+
+
+@marvin.fn
+def get_locations(location_strings: list[str]) -> list[Location]:
+    """
+    Estimate geo locations from fuzzy input strings.
+
+    Args:
+        location_strings (list[str]): List of location strings.
+
+    Returns:
+        list[Location]: List of Location objects with extracted details and geo coordinates.
+
+    Format for each input string:
+        <type> <priority> <location> <description> <city> <number>
+
+    Example components:
+        <priority>    : Priority level (e.g., 'High', 'Medium', 'Low').
+        <description> : Brief incident/service description (e.g., 'Accident', 'Service Call').
+        <location>    : Could contain street name, postal code, and/or city name.
+        <date + time> : Indicates the time of the incident/service.
+
+    Note:
+        - Street names or postal codes are the most specific identifiers.
+        - Use the city name if no street name or postal code is found.
+    """
+
+
+if __name__ == "__main__":
+    location_strings = [
+        "High Accident 123 Main St, New York 2024-06-28 12:59:00",
+        "Medium Service Call Los Angeles 2024-06-28 13:00:00",
+        "Low Fire Incident 456 Elm St, 60614, Chicago 2024-06-28 13:01:00",
+        "High Medical Emergency Uptown, Boston 2024-06-28 13:02:00",
+        "Medium Road Block 789 Pine St, San Francisco 2024-06-28 13:03:00",
+    ]
+    locations = get_locations(location_strings)
+
+    for location in locations:
+        print(location)
+
+    # Output:
+    # fuzzy_input_name='High Accident 123 Main St, New York 2024-06-28 12:59:00' location_string_extracted='123 Main St, New York' latitude=40.7127753 longitude=-74.0059728
+    # fuzzy_input_name='Medium Service Call Los Angeles 2024-06-28 13:00:00' location_string_extracted='Los Angeles' latitude=34.0522342 longitude=-118.2436849
+    # fuzzy_input_name='Low Fire Incident 456 Elm St, 60614, Chicago 2024-06-28 13:01:00' location_string_extracted='456 Elm St, 60614, Chicago' latitude=41.8493 longitude=-87.648
+    # fuzzy_input_name='High Medical Emergency Uptown, Boston 2024-06-28 13:02:00' location_string_extracted='Uptown, Boston' latitude=42.3600825 longitude=-71.0588801
+    # fuzzy_input_name='Medium Road Block 789 Pine St, San Francisco 2024-06-28 13:03:00' location_string_extracted='789 Pine St, San Francisco' latitude=37.7749295 longitude=-122.4194155
+```
+
+#marvinai