Function Libraries
The following library functions can be imported for additional functionality:
Math
Functions for advanced mathematics under the math package.
| Function Name | Description |
|---|---|
| cbrt(double) | Calculates the cube root of a number. For example, cbrt(27.0) returns 3.0, which is the cube root of 27.0. |
| copy_sign(double, double) | Returns the first argument with the sign of the second argument. For example, copy_sign(2.0, -3.0) returns -2.0. |
| expm1(double) | Calculates e^x - 1 with better precision for small values. For example, expm1(0.0) returns 0.0, as e^0 - 1 = 0. |
| hypot(double, double) | Computes sqrt(x² + y²) without intermediate overflow or underflow. For example, hypot(3.0, 4.0) returns 5.0, which is the hypotenuse of a 3-4-5 triangle. |
| log1p(double) | Computes the natural logarithm of 1 + x (log(1 + x)) accurately for small x. For example, log1p(0.0) returns 0.0 as log(1 + 0) = 0. |
| next_after(double, double) | Returns the next floating-point number towards the direction of the second argument. For example, next_after(1.0, 2.0) returns the next representable number after 1.0. |
| scalb(double, bigint) | Multiplies a floating-point number by 2 raised to the power of an integer. For example, scalb(1.0, 3) returns 8.0 as 1.0 * 2^3 = 8.0. |
| ulp(double) | Returns the size of the unit in the last place (ULP) of the argument. For example, ulp(1.0) returns the ULP of 1.0. |
| binomial_distribution(bigint, double, bigint) | Calculates the probability of obtaining a number of successes in a fixed number of trials for a binomial distribution. For example, binomial_distribution(10, 0.5, 5) returns the probability of 5 successes out of 10 trials with a 50% success rate. |
| exponential_distribution(double, double) | Evaluates the probability density or cumulative distribution of an exponential distribution. For example, exponential_distribution(1.0, 2.0) returns the exponential distribution's probability for a given rate and time. |
| normal_distribution(double, double, double) | Evaluates the cumulative distribution function for a normal (Gaussian) distribution. For example, normal_distribution(0.0, 1.0, 1.0) returns the probability for a standard normal distribution with mean 0 and standard deviation 1. |
| poisson_distribution(double, bigint) | Evaluates the probability mass function of a Poisson-distributed random variable. For example, poisson_distribution(1.0, 5) returns the probability of observing 5 events when the average event rate is 1.0. |
OpenAI
OpenAI API function under the openai package.
You can use these UDFs in your SQRL scripts to perform tasks such as text completion, extraction, and embedding, leveraging the power of OpenAI models.
For each function it is required to set the OPENAI_API_KEY environment variable with your OpenAI API key.
| Function Name | Description |
|---|---|
| completions(String prompt, String model_name) | Generates a completion for the given prompt using the specified OpenAI model. For example, completions('What is AI?', 'gpt-4o') returns a possible response to the prompt. |
| completions(String prompt, String model_name, Integer maxOutputTokens) | Generates a completion for the given prompt using the specified OpenAI model, with an upper limit on the number of output tokens. For example, completions('What is AI?', 'gpt-4o', 100) returns a possible response to the prompt, limited to 100 characters. |
| completions(String prompt, String model_name, Integer maxOutputTokens, Double temperature) | Generates a completion for the given prompt using the specified OpenAI model, with an upper limit on the number of output tokens and a specified temperature. For example, completions('What is AI?', 'gpt-4o', 100, 0.5) returns a possible response to the prompt, limited to 100 characters and weighted by a temperature of 0.5. |
| completions(String prompt, String model_name, Integer maxOutputTokens, Double temperature, Double topP) | Generates a completion for the given prompt using the specified OpenAI model, with an upper limit on the number of output tokens, a specified temperature, and a specified top-p value. For example, completions('What is AI?', 'gpt-4o', 100, 0.5, 0.9) returns a possible response to the prompt, limited to 100 characters, weighted by a temperature of 0.5, and with a top-p value of 0.9. |
| extract_json(String prompt, String model_name) | Extracts JSON data from the given prompt using the specified OpenAI model. For example, extract_json('What is AI?', 'gpt-4o') returns any relevant JSON data for the prompt. |
| extract_json(String prompt, String model_name, Double temperature) | Extracts JSON data from the given prompt using the specified OpenAI model and a specified temperature. For example, extract_json('What is AI?', 'gpt-4o', 0.5) returns any relevant JSON data for the prompt, weighted by a temperature of 0.5. |
| extract_json(String prompt, String model_name, Double temperature, Double topP) | Extracts JSON data from the given prompt using the specified OpenAI model, with a specified temperature and top-p value. For example, extract_json('What is AI?', 'gpt-4o', 0.5, 0.9) returns any relevant JSON data for the prompt, weighted by a temperature of 0.5 and with a top-p value of 0.9. |
| vector_embedd(String text, String model_name) | Embeds the given text into a vector using the specified OpenAI model. For example, vector_embedd('What is AI?', 'text-embedding-ada-002') returns a vector representation of the text. |