SummarizeContent

Summarize the contents of a file or URL

Chat Example

!SummarizeContent

ExampleData/document.nb

The document contains a mix of text, mathematical expressions, graphics, and dynamic objects.

In the text section, it simply contains a line of text and a subsection containing a set of words.

In the expressions section, three mathematical expressions are evaluated: addition of two numbers, the numerical approximation of the Golden Ratio to 80 decimal places, and a nested style function.

In the graphics section, two 3D graphics are created: a sphere and a plot of the absolute value function.

In the dynamic objects section, two dynamic objects are created: a slider and a display for the value of the slider.

!SummarizeContent

https://writings.stephenwolfram.com/2023/05/the-new-world-of-llm-functions-integrating-llm-technology-into-the-wolfram-language/

The blog post by Stephen Wolfram discusses integrating LLM (large language model) technology into the Wolfram Language. The author introduces the new LLMFunction, which allows users to "package up" LLM functionality and use it as a component in other operations. This functionality will be built into the upcoming version 13.3 of the Wolfram Language, but can also be installed in the current version (13.2).

In addition, the post explores how LLMFunction can be used, providing examples of its application in rewriting sentences in active voice, finding word analogies, and using everyday knowledge and creativity. The LLMFunction is a symbolic object that can be applied to an argument to get a result. The author also explains how one can map LLMFunction over a list and progressively nest an LLMFunction.

The post also covers the use of LLM functions in extracting structured data from text and turning it into a computable form. For instance, one can ask quantity questions based on text and get back structured data. In addition, the author introduces the LLMExampleFunction, which allows for learning from examples, making it easy for the LLM to apply what it has learned.

In conclusion, the integration of LLMs into the Wolfram Language provides a new world of possibilities, allowing users to leverage the capabilities of these models in a wide range of applications.

In[]:=

CloudExport

all countries, dependencies, and territories

COUNTRIES

[{"Name","Population"}]/.q_QuantityQuantityMagnitude[q],"CSV","countriespopulation.csv"

Out[]=

CloudObject[

https://www.wolframcloud.com/obj/rhennigan/countriespopulation.csv

]

!SummarizeContent

The given URL or file contains a list of countries along with their respective populations. The data is presented in the format "Country", "Population". The populations are presented as numerical values, except for Bouvet Island, which is listed as "Missing[""NotAvailable""]". The country listed with the highest population is China with 1425700000 and the country with the lowest listed population is the Pitcairn Islands with 56.

Programmatic Access

Retrieve the prompt:
LLMPrompt["SummarizeContent"]

Examples

Summarize a web page:

In[1]:=

LLMResourceFunction["SummarizeContent"]["https://www.wolfram.com"]

Out[1]=

Wolfram is a technology company that specializes in computation, data, and decision-making solutions. They offer various products and services, including Wolfram|One, Mathematica, Wolfram|Alpha Notebook Edition, Finance Platform, System Modeler, and Wolfram Player. Wolfram's technologies include the Wolfram Language, Wolfram Notebooks, Wolfram Data Framework, Wolfram Cloud, Wolfram Engine, and Wolfram Universal Deployment System. They provide solutions for engineering, R&D, finance, statistics, business analysis, software development, and education, among other fields. Additionally, they offer learning resources and support services, including Wolfram Language documentation, webinars, training, and technical consulting.

Summarize a notebook:

In[2]:=

LLMResourceFunction["SummarizeContent"]["ExampleData/document.nb"]

Out[2]=

The document contains text and mathematical expressions. The text is divided into two sections: "Text" and "Expressions". The "Text" section contains a line of text and a subsection with the words "one two three" and "alpha beta gamma delta epsilon". The "Expressions" section contains several mathematical expressions, including a nested style function, a 3D graphic of a sphere, a plot of the absolute value of x from -2 to 2, and a slider.

Summarize a

PDF

file:

In[3]:=

LLMResourceFunction["SummarizeContent"]["https://exampledata.wolfram.com/article.pdf"]

Out[3]=

The paper titled "Cut to Fit: Tailoring the Partitioning to the Computation" by Iacovos Kolokasis and Polyvios Pratikakis discusses optimization of partitioning strategies for social graph analytics applications. The authors evaluate six graph partitioning algorithms on six social graphs using four standard graph algorithms and measure their performance based on five partitioning metrics. They identify communication cost as the best predictor of performance for most analytics computations. The authors also conclude that the optimal partitioning strategy depends on the properties of the graph dataset, the analytics computation, and the number of partitions. These insights can significantly impact the performance of large graph analytics computations and help optimize partitioning strategies for specific datasets and computations.

Scope
(3)

Summarize a

URL

object:

In[1]:=

LLMResourceFunction["SummarizeContent"][URL["https://www.wolfram.com"]]

Out[1]=

Wolfram is a technology company that offers a range of products and services including Wolfram|One, Mathematica, Wolfram|Alpha Notebook Edition, Finance Platform, System Modeler, Wolfram Player, Wolfram Engine, and WolframScript. They also provide enterprise solutions, corporate consulting, technical consulting, and Wolfram|Alpha business solutions. Their technologies include the Wolfram Language, Wolfram Notebooks, Wolfram Data Framework, Wolfram Cloud, Wolfram Engine, and Wolfram Universal Deployment System. They offer solutions for engineering, R&D, education, finance, statistics, business analysis, software & web development, and sciences. They also provide learning and support resources, and have initiatives for startups, demonstrations, and summer programs. The website also features a blog and information about the company's background and career opportunities.

Summarize a

CloudObject

In[2]:=

CloudExport["Hello world","String"]

Out[2]=

CloudObject[

https://www.wolframcloud.com/obj/b7078af2-a9e0-48fb-8298-1571951d8f3a

]

In[3]:=

LLMResourceFunction["SummarizeContent"][%]

Out[3]=

The content of the given URL is a simple text message that says "Hello world".

Summarize a

LocalObject

In[4]:=

Export[LocalObject[],"Hello world","String"]

Out[4]=

LocalObject[file:///C:/Users/rhennigan/AppData/Roaming/Wolfram/Objects/25e42b39-9682-4ac1-9d53-0c3de30c1ffd]

In[5]:=

LLMResourceFunction["SummarizeContent"][%]

Out[5]=

The file contains the text "Hello world".

Properties and Relations
(1)



Possible Issues
(1)



Prompt Source

Definition Notebook »

Publisher Information

Contributed by: Richard Hennigan (Wolfram Research)

Wolfram Prompt Repository(Under Construction)

SummarizeContent

Chat Example

Programmatic Access

Examples

Scope
(3)

Properties and Relations
(1)



Possible Issues
(1)



Prompt Source

See Also

Publisher Information

SummarizeContent

Chat Example

Programmatic Access

Examples

Scope (3)

Properties and Relations (1) 

Possible Issues (1) 

Prompt Source

See Also

Publisher Information

Scope
(3)

Properties and Relations
(1)



Possible Issues
(1)

