| Type: | Package |
| Title: | A Coding Assistant using Perplexity's Large Language Models |
| Version: | 0.0.3 |
| Maintainer: | Gabriel Kaiser <quantresearch.gk@gmail.com> |
| Description: | A coding assistant using Perplexity's Large Language Models https://www.perplexity.ai/ API. A set of functions and 'RStudio' add-ins that aim to help R developers. |
| License: | GPL (≥ 3) |
| URL: | https://github.com/GabrielKaiserQFin/perplexR |
| BugReports: | https://github.com/GabrielKaiserQFin/perplexR/issues |
| Imports: | clipr, httr, jsonlite, miniUI, rstudioapi, shiny, utils |
| Encoding: | UTF-8 |
| Language: | en-US |
| RoxygenNote: | 7.2.3 |
| NeedsCompilation: | no |
| Packaged: | 2024-03-28 09:58:24 UTC; Gabriel |
| Author: | Gabriel Kaiser [aut, cre] |
| Repository: | CRAN |
| Date/Publication: | 2024-03-29 20:50:02 UTC |
perplexR: A Coding Assistant using Perplexity's Large Language Models
Description
A coding assistant using Perplexity's Large Language Models https://www.perplexity.ai/ API. A set of functions and 'RStudio' add-ins that aim to help R developers.
Author(s)
Maintainer: Gabriel Kaiser quantresearch.gk@gmail.com
See Also
Useful links:
Report bugs at https://github.com/GabrielKaiserQFin/perplexR/issues
Get Large Language Model Completions Endpoint
Description
Get Large Language Model Completions Endpoint
Usage
API_Request(
prompt,
PERPLEXITY_API_KEY = PERPLEXITY_API_KEY,
modelSelection = modelSelection,
systemRole = systemRole,
maxTokens = maxTokens,
temperature = temperature,
top_p = top_p,
top_k = top_k,
presence_penalty = presence_penalty,
frequency_penalty = frequency_penalty,
proxy = proxy
)
Arguments
prompt |
The prompt to generate completions for. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Ask Large Language Model
Description
Note: See also clearChatSession.
Usage
AskMe(
question,
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
question |
The question to ask Large Language Model. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
AskMe("What do you think about Large language models?")
## End(Not run)
Large Language Model: Annotate code
Description
Large Language Model: Annotate code
Usage
annotateCode(
code = clipr::read_clip(allow_non_interactive = TRUE),
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
The code to be commented by Large Language Model. If not provided, it will use what's copied on the clipboard. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
annotateCode("z <- function(x) scale(x)^2")
## End(Not run)
Large Language Model: Create Unit Tests
Description
Create {testthat} test cases for the code.
Usage
buildUnitTests(
code = clipr::read_clip(allow_non_interactive = TRUE),
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
The code for which to create unit tests by Large Language Model. If not provided, it will use what's copied on the clipboard. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
buildUnitTests("squared_numbers <- function(numbers) {\n numbers ^ 2\n}")
## End(Not run)
Large Language Model: Clarify Code
Description
Large Language Model: Clarify Code
Usage
clarifyCode(
code = clipr::read_clip(allow_non_interactive = TRUE),
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
The code to be explained by Large Language Model. If not provided, it will use what's copied on the clipboard. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
clarifyCode("z <- function(x) scale(x)^2")
## End(Not run)
Large Language Model: Find Issues in Code
Description
Large Language Model: Find Issues in Code
Usage
debugCode(
code = clipr::read_clip(allow_non_interactive = TRUE),
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
The code to be analyzed by Large Language Model. If not provided, it will use what's copied on the clipboard. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
debugCode("z <- function(x) scale(x)2")
## End(Not run)
Large Language Model: Code Documentation (roxygen2 style)
Description
Large Language Model: Code Documentation (roxygen2 style)
Usage
documentCode(
code = clipr::read_clip(allow_non_interactive = TRUE),
inLineDocumentation = "roxygen2",
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
The code to be documented by Large Language Model. If not provided, it will use what's copied on the clipboard. |
inLineDocumentation |
Formatting style of In-Line Documentation. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
documentCode("z <- function(x) scale(x)^2")
## End(Not run)
Run a Large Language Model as RStudio add-in
Description
Run a Large Language Model as RStudio add-in
Usage
execAddin(FUN)
Arguments
FUN |
The function to be executed. |
Ask Large Language Model
Description
Opens an interactive chat session with Large Language Model
Usage
execAddin_AskMe()
Large Language Model: Finish code
Description
Large Language Model: Finish code
Usage
finishCode(
code = clipr::read_clip(allow_non_interactive = TRUE),
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
The code to be completed by Large Language Model. If not provided, it will use what's copied on the clipboard. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
finishCode("# A function to square each element of a vector\nsquare_each <- function(")
## End(Not run)
Large Language Model: Create a Function or Variable Name
Description
Large Language Model: Create a Function or Variable Name
Usage
namingGenie(
code = clipr::read_clip(allow_non_interactive = TRUE),
namingConvention = "camelCase",
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
The code for which to give a variable name to its result. If not provided, it will use what's copied on the clipboard. |
namingConvention |
Naming convention. Default is camelCase. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
namingGenie("sapply(1:10, function(i) i ** 2)")
## End(Not run)
Large Language Model: Optimize Code
Description
Large Language Model: Optimize Code
Usage
optimizeCode(
code = clipr::read_clip(allow_non_interactive = TRUE),
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
The code to be optimized by Large Language Model. If not provided, it will use what's copied on the clipboard. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
optimizeCode("z <- function(x) scale(x)^2")
## End(Not run)
Parse Perplexity API Response
Description
Takes the raw response from the Perplexity API and extracts the text content from it.
Usage
responseParser(raw)
Arguments
raw |
The raw object returned by the Perplexity API. |
Value
Returns a character vector containing the text content of the response.
responseReturn
Description
responseReturn
Usage
responseReturn(raw)
Arguments
raw |
the chatresponse to return |
Value
A character value with the response generated by Large Language Model.
Large Language Model: Rewrite Text
Description
Large Language Model: Rewrite Text
Usage
rewriteText(
text = clipr::read_clip(allow_non_interactive = TRUE),
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
text |
The text to be rewritten by Large Language Model. If not provided, it will use what's copied on the clipboard. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
rewriteText("Dear Recipient, I hope this message finds you well.")
## End(Not run)
Translate Code from One Language to Another
Description
This function takes a snippet of code and translates it from one programming language to another using Perplexity API. The default behavior is to read the code from the clipboard and translate from R to Python.
Usage
translateCode(
code = clipr::read_clip(allow_non_interactive = TRUE),
from = "R",
to = "Python",
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant with extensive programming skills.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
code |
A string containing the code to be translated. If not provided, the function will attempt to read from the clipboard. |
from |
The language of the input code. Default is "R". |
to |
The target language for translation. Default is "Python". |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant with extensive knowledge of R programming." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A string containing the translated code.
Examples
## Not run:
translateCode("your R code here", from = "R", to = "Python")
## End(Not run)
Large Language Model: Translate Text
Description
Large Language Model: Translate Text
Usage
translateText(
text = clipr::read_clip(allow_non_interactive = TRUE),
toLanguage = "German",
PERPLEXITY_API_KEY = Sys.getenv("PERPLEXITY_API_KEY"),
modelSelection = c("mistral-7b-instruct", "mixtral-8x7b-instruct",
"codellama-70b-instruct", "sonar-small-chat", "sonar-small-online",
"sonar-medium-chat", "sonar-medium-online"),
systemRole = "You are a helpful assistant.",
maxTokens = 265,
temperature = 1,
top_p = NULL,
top_k = 100,
presence_penalty = 0,
frequency_penalty = NULL,
proxy = NULL
)
Arguments
text |
The text to be translated by Large Language Model. If not provided, it will use what's copied on the clipboard. |
toLanguage |
The language to be translated into. |
PERPLEXITY_API_KEY |
PERPLEXITY API key. |
modelSelection |
model choice. Default is mistral-7b-instruct. |
systemRole |
Role for model. Default is: "You are a helpful assistant." |
maxTokens |
The maximum integer of completion tokens returned by API. |
temperature |
The amount of randomness in the response, valued between 0 inclusive and 2 exclusive. Higher values are more random, and lower values are more deterministic. Set either temperature or top_p. |
top_p |
Nucleus sampling threshold, valued between 0 and 1 inclusive. |
top_k |
The number of tokens to keep for highest top-k filtering, specified as an integer between 0 and 2048 inclusive. If set to 0, top-k filtering is disabled. |
presence_penalty |
A value between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics. Incompatible with frequency_penalty. |
frequency_penalty |
A multiplicative penalty greater than 0. Values greater than 1.0 penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim. A value of 1.0 means no penalty. |
proxy |
Default value is NULL. |
Value
A character value with the response generated by Large Language Model.
Examples
## Not run:
translateText("Dear Recipient, I hope this message finds you well.")
## End(Not run)