[BUG] hackernews demo fails on merge branch

[lmeyerov]

On http://localhost/notebook/lab/tree/demos/ai/Introduction/Ask-HackerNews-Demo.ipynb:

ile /opt/conda/envs/rapids/lib/python3.8/site-packages/graphistry/feature_utils.py:652, in impute_and_scale_df(df, use_scaler, impute, n_quantiles, output_distribution, quantile_range, n_bins, encode, strategy, keep_n_decimals)
    629 def impute_and_scale_df(
    630     df: pd.DataFrame,
    631     use_scaler: str = "robust",
   (...)
    639     keep_n_decimals: int = 5,
    640 ) -> Tuple[pd.DataFrame, Pipeline]:
    642     transformer = get_preprocessing_pipeline(
    643         impute=impute,
    644         use_scaler=use_scaler,
   (...)
    650         strategy=strategy,
    651     )
--> 652     res = fit_pipeline(df, transformer, keep_n_decimals=keep_n_decimals)
    654     return res, transformer

File /opt/conda/envs/rapids/lib/python3.8/site-packages/graphistry/feature_utils.py:622, in fit_pipeline(X, transformer, keep_n_decimals)
    619 columns = X.columns
    620 index = X.index
--> 622 X = transformer.fit_transform(X)
    623 if keep_n_decimals:
    624     X = np.round(X, decimals=keep_n_decimals)  #  type: ignore  # noqa

File /opt/conda/envs/rapids/lib/python3.8/site-packages/sklearn/pipeline.py:437, in Pipeline.fit_transform(self, X, y, **fit_params)
    410 """Fit the model and transform with the final estimator.
    411 
    412 Fits all the transformers one after the other and transform the
   (...)
    434     Transformed samples.
    435 """
    436 fit_params_steps = self._check_fit_params(**fit_params)
--> 437 Xt = self._fit(X, y, **fit_params_steps)
    439 last_step = self._final_estimator
    440 with _print_elapsed_time("Pipeline", self._log_message(len(self.steps) - 1)):

File /opt/conda/envs/rapids/lib/python3.8/site-packages/sklearn/pipeline.py:339, in Pipeline._fit(self, X, y, **fit_params_steps)
    336 def _fit(self, X, y=None, **fit_params_steps):
    337     # shallow copy of steps - this should really be steps_
    338     self.steps = list(self.steps)
--> 339     self._validate_steps()
    340     # Setup the memory
    341     memory = check_memory(self.memory)

File /opt/conda/envs/rapids/lib/python3.8/site-packages/sklearn/pipeline.py:243, in Pipeline._validate_steps(self)
    237 # We allow last estimator to be None as an identity transformation
    238 if (
    239     estimator is not None
    240     and estimator != "passthrough"
    241     and not hasattr(estimator, "fit")
    242 ):
--> 243     raise TypeError(
    244         "Last step of Pipeline should implement fit "
    245         "or be the string 'passthrough'. "
    246         "'%s' (type %s) doesn't" % (estimator, type(estimator))
    247     )

TypeError: Last step of Pipeline should implement fit or be the string 'passthrough'. '<function identity at 0x7fc7b4870430>' (type <class 'function'>) doesn't

[lmeyerov]

as w/ other semantic search, now also getting:

File /opt/conda/envs/rapids/lib/python3.8/site-packages/pandas/io/excel/_base.py:1656, in ExcelFile.__init__(self, path_or_buffer, engine, storage_options)
   1652     ext = inspect_excel_format(
   1653         content_or_path=path_or_buffer, storage_options=storage_options
   1654     )
   1655     if ext is None:
-> 1656         raise ValueError(
   1657             "Excel file format cannot be determined, you must specify "
   1658             "an engine manually."
   1659         )
   1661 engine = config.get_option(f"io.excel.{ext}.reader", silent=True)
   1662 if engine == "auto":

ValueError: Excel file format cannot be determined, you must specify an engine manually.

so semantic search may have an issue?

[lmeyerov]

@tanmoyio search still problematic on hn demo even for 0.29.1 (g distro so w/ rapids):

[lmeyerov]

I'm getting a lot of:

gr = g2.search_graph('How to create deep learning models', thresh=15, top_n=50, scale=0.25, broader=False) 
gr.plot()

=>

**No results found due to empty DataFrame, returning original graph

[lmeyerov]

I wonder if this is parameterization based on cpu vs gpu, or something deeper?