Text-based representations of location data for large language model-based item identification
An online system generates text-based representations of various types of data for processing using a large language model. The online system extracts location data from a map of a source location and converts the location data into a text-based representation of the location data. The online system receives a set of item identifiers from a client device of a user and generates an LLM prompt based on the set of item identifiers and the text-based representations of the location data. The online system receives a response from the LLM and parses the response for a text-based description of related items. The online system maps the text-based description of the related items to item identifiers and transmits a notification to the client device that includes item data associated with the related items.
1 . A method comprising:
extracting location data from a map of a source location, wherein the location data describes locations of a plurality of items within the source location;
converting the location data into a text-based representation of the location data, wherein the text-based representation comprises text describing the location data;
receiving a set of item identifiers from a client device associated with a user, wherein each of the item identifiers corresponds to an item of the plurality of items;
generating a prompt for input to a large language model, wherein the prompt comprises item data describing items corresponding to the set of item identifiers, the text-based representation of the location data, and instructions to identify a set of related items of the plurality of items based on the item data and the text-based representation of the location data;
receiving a response from the large language model;
parsing the response to extract a text-based description of the set of related items;
mapping the text-based description of the set of related items to an identifier for each of the set of related items; and
transmitting a notification to the client device for display to the user based on the identifiers for the set of related items, wherein the notification comprises item data associated with each of the set of related items and locations of the set of related items within the source location.
2 . The method of claim 1 , further comprising:
receiving device location data captured by a sensor coupled to the client device, wherein the device location data describes a location of the client device;
generating a text-based representation of the device location data; and
generating the prompt for input to the large language model based on the text-based representation of the device location data.
3 . The method of claim 2 , further comprising:
identifying the source location for items corresponding to the set item identifies
based on the received device location data.
4 . The method of claim 1 , wherein converting the location data into a text-based representation of the location data comprises:
applying a template or rules-based logic to the location data.
5 . The method of claim 1 , wherein generating the prompt comprises:
accessing an item database storing item data for the plurality of items to retrieve the item data; and
generating the prompt based on the retrieved item data.
6 . The method of claim 1 , wherein the prompt comprises instructions for the large language model to generate a route for the user to collect items corresponding to the set of item identifiers, and wherein the notification comprises the route for the user to collect items corresponding to the set of item identifiers.
7 . The method of claim 1 , wherein the map indicates locations of categories of items within the source location.
8 . The method of claim 1 , wherein the map indicates a location for each item of the plurality of items.
9 . The method of claim 1 , wherein the prompt for input to the large language model comprises instructions to generate the text-based description of the set of related items according to a specific format and wherein the response is parsed based on the specific format.
10 . The method of claim 1 , wherein the text-based description of the set of related items comprises a description of a generic item for each of the set of related items.
11 . A non-transitory computer-readable medium storing instructions that, when executed, cause a processor to perform operations comprising:
extracting location data from a map of a source location, wherein the location data describes locations of a plurality of items within the source location;
converting the location data into a text-based representation of the location data, wherein the text-based representation comprises text describing the location data;
receiving a set of item identifiers from a client device associated with a user, wherein each of the item identifiers corresponds to an item of the plurality of items;
generating a prompt for input to a large language model, wherein the prompt comprises item data describing items corresponding to the set of item identifiers, the text-based representation of the location data, and instructions to identify a set of related items of the plurality of items based on the item data and the text-based representation of the location data;
receiving a response from the large language model;
parsing the response to extract a text-based description of the set of related items;
mapping the text-based description of the set of related items to an identifier for each of the set of related items; and
transmitting a notification to the client device for display to the user based on the identifiers for the set of related items, wherein the notification comprises item data associated with each of the set of related items and locations of the set of related items within the source location.
12 . The non-transitory computer-readable medium of claim 11 , the operations further comprising:
receiving device location data captured by a sensor coupled to the client device, wherein the device location data describes a location of the client device;
generating a text-based representation of the device location data; and
generating the prompt for input to the large language model based on the text-based representation of the device location data.
13 . The non-transitory computer-readable medium of claim 12 , the operations further comprising:
identifying the source location for the set of items based on the received device location data.
14 . The non-transitory computer-readable medium of claim 11 , wherein converting the location data into a text-based representation of the location data comprises:
applying a template or rules-based logic to the location data.
15 . The non-transitory computer-readable medium of claim 11 , wherein generating the prompt comprises:
accessing an item database storing item data for the plurality of items to retrieve the item data; and
generating the prompt based on the retrieved item data.
16 . The non-transitory computer-readable medium of claim 11 , wherein the prompt comprises instructions for the large language model to generate a route for the user to collect items corresponding to the set of item identifiers, and wherein the notification comprises the route for the user to collect items corresponding to the set of item identifiers.
17 . The non-transitory computer-readable medium of claim 11 , wherein the map indicates locations of categories of items within the source location.
18 . The non-transitory computer-readable medium of claim 11 , wherein the map indicates a location for each item of the plurality of items.
19 . The non-transitory computer-readable medium of claim 11 , wherein the prompt for input to the large language model comprises instructions to generate the text-based description of the set of related items according to a specific format and wherein the response is parsed based on the specific format.
20 . A system comprising:
a processor; and
a non-transitory computer-readable medium storing instructions that, when executed, cause a processor to perform operations comprising:
extracting location data from a map of a source location, wherein the location data describes locations of a plurality of items within the source location;
converting the location data into a text-based representation of the location data, wherein the text-based representation comprises text describing the location data;
receiving a set of item identifiers from a client device associated with a user, wherein each of the item identifiers corresponds to an item of the plurality of items;
generating a prompt for input to a large language model, wherein the prompt comprises item data describing items corresponding to the set of item identifiers, the text-based representation of the location data, and instructions to identify a set of related items of the plurality of items based on the item data and the text-based representation of the location data;
receiving a response from the large language model;
parsing the response to extract a text-based description of the set of related items;
mapping the text-based description of the set of related items to an identifier for each of the set of related items; and
transmitting a notification to the client device for display to the user based on the identifiers for the set of related items, wherein the notification comprises item data associated with each of the set of related items and locations of the set of related items within the source location.