Package 'GetTDData'

Returns metadata and information about Tesouro Direto bond types

Description

Provides metadata for Brazilian government bond types, including their official names, indexers (e.g., IPCA, SELIC, Fixed rate), payment frequency, and description.

Usage

get_asset_info(asset_codes = NULL)

Arguments

asset_codes

Optional character vector of asset codes (e.g. 'LTN', 'NTN-B'). If 'NULL', returns info for all assets.

Value

A tibble with asset metadata.

Examples

get_asset_info()
get_asset_info("LTN")

---

Returns the cache directory path

Description

Returns the cache directory path

Usage

get_cache_folder(persistent = getOption("GetTDData.persistent_cache", FALSE))

Arguments

persistent

Logical. If 'TRUE', uses a user-persistent directory via 'tools::R_user_dir'. If 'FALSE', defaults to session temporary directory. Can also be set globally via 'options(GetTDData.persistent_cache = TRUE)'. Defaults to 'FALSE'.

Value

A character string representing the folder path.

Examples

get_cache_folder()

---

Returns the available asset names at the Tesouro Direto (TD) website

Description

Returns the available asset names at the Tesouro Direto (TD) website

Usage

get_td_names(online = FALSE)

Arguments

online

Logical. If 'TRUE', attempts to query the Tesouro Direto CDN server for live asset names. Defaults to 'FALSE'.

Value

A character vector of names.

Examples

get_td_names()

---

Gets the current yield curve

Description

Downloads and parses information about the current Brazilian yield curve from Anbima.

Usage

get_yield_curve()

get.yield.curve()

Value

A data frame with information about the yield curve.

Examples

## Not run: 
df_yield <- get_yield_curve()
str(df_yield)

## End(Not run)

---

Run a GetTDData MCP server

Description

Starts a Model Context Protocol (MCP) server that exposes a curated set of 'GetTDData' functions as tools for LLM clients such as OpenCode, Claude Desktop, Claude Code, or Positron Assistant. The suggested package 'mcptools' is required.

Usage

mcp_server(
  tools = mcp_tools(),
  type = c("stdio", "http"),
  session_tools = FALSE,
  ...
)

Arguments

tools	Either a list of 'ellmer::tool()' objects, or the path to an R file that evaluates to such a list. Defaults to the tool definitions shipped with the package (see [mcp_tools()]).
type	Transport type: '"stdio"' (default) for clients that launch the process, or '"http"' for an HTTP server.
session_tools	Logical. Whether to expose mcptools' built-in R session discovery tools ('list_r_sessions', 'select_r_session'). Defaults to 'FALSE' because GetTDData tools run self-contained in the server process.
...	Further arguments passed on to 'mcptools::mcp_server()', such as 'host' or 'port'.

Details

Register the server with an MCP client by running 'Rscript -e "GetTDData::mcp_server()"'. For example, in OpenCode:

“'json

"mcp": "servers": "GetTDData": "type": "local", "command": ["Rscript", "-e", "GetTDData::mcp_server()"]

“'

Value

Called for its side effect of blocking the R process while the server runs. It does not return.

Examples

## Not run: 
# Requires the mcptools, ellmer, and jsonlite packages
GetTDData::mcp_server()

## End(Not run)

---

Build the Model Context Protocol (MCP) tools exposed by GetTDData

Description

Sources the MCP tool definitions shipped with the package ('inst/mcp/tools.R') and returns them as a list of 'ellmer::tool()' objects. These tools wrap a curated set of 'GetTDData' functions so that MCP-enabled LLM clients can fetch Brazilian government bond data.

Usage

mcp_tools(path = system.file("mcp", "tools.R", package = "GetTDData"))

Arguments

path	Path to an R file that, when sourced, evaluates to a list of 'ellmer::tool()' objects. Defaults to the definitions shipped with the package. This argument is mainly useful for testing custom tool sets.

Details

The suggested packages 'ellmer' and 'jsonlite' are required. The tool set currently exposes: 'get_td_data()', 'get_td_current()', 'get_yield_curve()', 'get_asset_info()', and 'get_td_names()'.

Value

A named list of 'ellmer::tool()' objects.

Examples

## Not run: 
# Requires the ellmer and jsonlite packages
tools <- mcp_tools()
names(tools)

## End(Not run)

---

Plot Tesouro Direto Time Series Data

Description

Plots price or yield time series trajectories for Tesouro Direto assets downloaded using td_get.

Usage

plot_td_series(df_td, type = c("price", "yield"))

Arguments

df_td	A data frame or tibble returned by td_get.
type	Character string indicating what to plot: "price" (price_bid) or "yield" (yield_bid). Defaults to "price".

Value

A ggplot object.

Examples

## Not run: 
df_td <- td_get("LTN", 2020, 2022)
plot_td_series(df_td, type = "price")

## End(Not run)

---

Plot the ANBIMA Yield Curve

Description

Plots the yield curve data frame downloaded using get_yield_curve.

Usage

plot_yield_curve(df_yc)

Arguments

df_yc

A data frame or tibble returned by get_yield_curve.

Value

A ggplot object.

Examples

## Not run: 
df_yc <- get_yield_curve()
plot_yield_curve(df_yc)

## End(Not run)

---

Downloads data for Brazilian government bonds directly from the website

Description

'r lifecycle::badge("deprecated")'

Usage

td_get(
  asset_codes = "LTN",
  first_year = 2005,
  last_year = as.numeric(format(Sys.Date(), "%Y")),
  dl_folder = get_cache_folder()
)

Arguments

asset_codes	A character vector identifying the assets (one or more) in the names of the Excel files (e.g., 'LTN'). If 'NULL', downloads all available assets.
first_year	The first year of data (minimum of 2005).
last_year	The last year of data.
dl_folder	Path of the folder to save Excel files from Tesouro Direto (will create if it does not exist). Defaults to a session-temporary directory. To avoid redownloading files across different R sessions, you can pass a persistent path (e.g., a local folder path, or using tools::R_user_dir("GetTDData", which = "cache")).

Details

This function looks into the Tesouro Direto website (<https://www.tesourodireto.com.br/>) and downloads all files containing prices and yields of government bonds. You can use the input 'asset_codes' to restrict the downloads to specific bonds.

Due to the closure of the Tesouro Direto data application in August 2026, 'td_get()' is deprecated and resulting data will only go as far as August 2026. Users should use 'td_get2()', a new function that will be developed soon.

Value

A data frame containing the asset data (prices and yields).

Examples

## Not run: 
df_td <- td_get("LTN", 2020, 2022)

## End(Not run)

---

Returns current TD prices and yields

Description

Fetches current prices and yields of Tesouro Direto (TD) assets by downloading the latest daily data available from the website.

Usage

td_get_current(asset_codes = NULL, dl_folder = get_cache_folder())

Arguments

asset_codes	A character vector identifying the assets (e.g., 'LTN', 'NTN-B'). If 'NULL', returns all available assets.
dl_folder	Path of the folder to save files. Defaults to a session-temporary directory.

Value

A tibble with current asset prices, yields, and maturity dates.

Examples

## Not run: 
df_current <- td_get_current()
head(df_current)

## End(Not run)

---

Downloads data for Brazilian government bonds from Tesouro Transparente (CKAN)

Description

This function downloads historical prices and yields of Brazilian government bonds directly from the Tesouro Transparente CKAN open data portal (<https://www.tesourotransparente.gov.br/ckan/dataset/taxas-dos-titulos-ofertados-pelo-tesouro-direto/>). It serves as the modern replacement for td_get, which relied on the legacy Tesouro Direto application shut down in August 2026.

Usage

td_get2(
  asset_codes = "LTN",
  first_year = 2005,
  last_year = as.numeric(format(Sys.Date(), "%Y")),
  dl_folder = get_cache_folder(),
  force_download = FALSE,
  dataset_url = get_default_ckan_url()
)

Arguments

asset_codes	A character vector identifying the assets to download (e.g., 'LTN', 'NTN-B', 'LFT', 'NTN-B Principal', 'NTN-F', 'NTN-C', 'NTN-B1', 'Educa+', 'RendA+'). You can also pass official names such as 'Tesouro Prefixado', 'Tesouro Selic', etc. If ‘NULL', downloads all available assets. Defaults to ’LTN'.
first_year	The first year of data (minimum of 2005). Defaults to 2005.
last_year	The last year of data. Defaults to the current year.
dl_folder	Path of the folder to save data files from Tesouro Transparente. Defaults to get_cache_folder().
force_download	Logical. If 'TRUE', forces redownloading the data even if a cached version from today is present. Defaults to 'FALSE'.
dataset_url	The URL to the CKAN dataset page or direct CSV file. Defaults to "https://www.tesourotransparente.gov.br/ckan/dataset/taxas-dos-titulos-ofertados-pelo-tesouro-direto/".

Value

A tibble (data frame) containing the asset data:

ref_date

Reference date (Date)

yield_bid

Annual purchase yield / rate in decimal format (numeric, e.g. 0.1183 for 11.83%)

price_bid

Unit purchase price in BRL (numeric)

asset_code

Standard bond identifier (character, e.g. 'LTN 010123')

matur_date

Maturity date of the bond (Date)

Examples

## Not run: 
df_td <- td_get2("LTN", 2020, 2022)
head(df_td)

## End(Not run)

---