See the fragment files in the changelog.d directory.
- Removed the older models from model shortcuts and docs. (#853)
- Fixed output validation to handle Python code block markers from llm responses. (#858)
- Changed prompt for
vizro-ai.plot
to provide more guidance for desired import format. (#854)
- VizroAI now allows any model with
langchain
structured output capabilities to be used, not justChatOpenAI
. (#646) - VizroAI now has a more flexible output when choosing
VizroAI.plot(...,return_elements=True)
. See Vizro-AI docs for all new options. (#646) - VizroAI now supports text-to-dashboard generation using
VizroAI.dashboard()
. To get started, visit the Vizro-AI docs.
- Removed the automatic display of chart explanation and insights in Jupyter. (#646)
- Changed the return type of
VizroAI.plot(...,return_elements=True)
fromPlotOutputs
dataclass to a pydantic model with more flexible methods. See Vizro-AI docs for more info. (#646)
- Removed argument
explain
from VizroAI.plot(). Usereturn_elements=True
instead. (#646)
- Added Docker image of
VizroAI.plot
UI for local use (#1177)
- Documented the step for instantiating
VizroAI
in text-to-dashboard pages (#685)
- Remove
_return_all_text
fromVizroAI
class (#518)
-
Add argument
return_elements
toVizroAI.plot()
. When it is set toTrue
, the return type will be changed to adataclass
containing the code string, figure object, business insights, and code explanation. (#488) -
Add functionality to generate dashboards from text. This feature is currently in Alpha and is not yet officially released (#651)
-
Disabled figure display upon variable assignment. To display outcome of
VizroAI.plot()
add.show()
. (#527) -
Stabilized
plot
performance by addressing several dataframe mutation issues. (#603)
- Removed
temperature
andmodel_name
arguments ofVizroAI
class. For current configuration options, visit the Vizro-AI docs (#423)
- Enable customization of LLM models provided to Vizro-AI class. (#423)
VizroAI.plot
now generates and returns aplotly.graph_objects.Figure
object. (#411). This facilitates the integration of Vizro-AI charts into thevizro
dashboard, visit Use Vizro-AI dynamically to return a fig object for details.
-
Add
max_debug_retry
parameter toVizroAI.plot
to allow users to determine the maximum number of debugging attempts desired. (#261) -
Enable automatic loading of environment variables in a
.env
file. (#270)
- Remove upper bound for
langchain
andopenai
dependencies. (#369)
- Remove the keyword
explain
from docs example explaining the_get_chart_code
function. (#256)
- Fix incompatibility with
pydantic>=2.0.0
(#189)
- Bump langchain version to 0.0.329, suggested by snyk (#204)
- Initial release of Vizro-AI package. Vizro-AI is a tool for generating data visualizations. (#138)