From 102cad584da9c4d11d75257856061e2e4a298436 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ru=20Ke=C3=AFn?=
 <3181182+alphasentaurii@users.noreply.github.com>
Date: Wed, 3 Apr 2024 20:22:55 -0400
Subject: [PATCH] Encoder handle unknown (#53)

* only create one encoding per unknown variable

* set default loglevel to INFO
---
 CHANGES.rst                     |  5 +++++
 spacekit/logger/log.py          |  2 +-
 spacekit/preprocessor/encode.py | 22 ++++++++++++----------
 3 files changed, 18 insertions(+), 11 deletions(-)

diff --git a/CHANGES.rst b/CHANGES.rst
index 4137e99..bfec3c4 100644
--- a/CHANGES.rst
+++ b/CHANGES.rst
@@ -10,6 +10,11 @@ new features
 
 - `extractor.scrape.FitsScraper.scrape_dataframe` method added for scraping Fits data from dataframe [#52]
 
+bug fixes
+---------
+
+- `preprocessor.encode.PairEncoder.handle_unknowns` create single new encoding value per unidentified variable [#53]
+
 
 1.0.1 (2024-04-03)
 ==================
diff --git a/spacekit/logger/log.py b/spacekit/logger/log.py
index faa7cba..ccbcada 100644
--- a/spacekit/logger/log.py
+++ b/spacekit/logger/log.py
@@ -68,7 +68,7 @@ def __init__(
         self.short_name = script_name
         self.console = console
         self.logfile = logfile
-        self.log_level = logging.DEBUG
+        self.log_level = logging.INFO
         self.console_log_output = console_log_output.lower()
         self.console_log_level = console_log_level.upper()
         self.console_log_color = console_log_color
diff --git a/spacekit/preprocessor/encode.py b/spacekit/preprocessor/encode.py
index 83c9da8..9765a1d 100644
--- a/spacekit/preprocessor/encode.py
+++ b/spacekit/preprocessor/encode.py
@@ -108,12 +108,13 @@ def inverse_pairs(self):
         return self.invpairs
 
     def handle_unknowns(self, unknowns):
-        self.log.warning(f"Found unknown values:\n {self.arr[unknowns]}")
-        add_encoding = max(list(self.keypairs.values())) + 1
+        uvals = np.unique(self.arr[unknowns])
+        self.log.warning(f"Found unknown values:\n {uvals}")
         try:
-            self.keypairs[self.arr[unknowns][0]] = add_encoding
-            self.classes_ = list(self.keypairs.keys())
-            self.log.info("Successfully added encoding for unknown values.")
+            for u in uvals:
+                add_encoding = max(list(self.keypairs.values())) + 1
+                self.keypairs[u] = add_encoding
+                self.classes_ = list(self.keypairs.keys())
         except Exception as e:
             self.log.error("Unable to add encoding for unknown value(s)", e)
 
@@ -236,16 +237,17 @@ def _encode_features(self):
                 "encoding_pairs attr must be instantiated with key-value pairs"
             )
             return
-        self.log.info("Encoding categorical features...")
+        self.log.debug("Encoding categorical features...")
         for col, name in self.encodings.items():
             keypairs = self.encoding_pairs[col]
             enc = PairEncoder()
             enc.fit_transform(self.df, keypairs, axiscol=col)
             self.df[name] = enc.transformed
-            self.log.debug(f"*** {col} --> {name} ***")
-            self.log.debug(
-                f"\n\nORIGINAL:\n{self.df[col].value_counts()}\n\nENCODED:\n{self.df[name].value_counts()}\n"
-            )
+            if self.verbose:
+                self.log.debug(f"*** {col} --> {name} ***")
+                self.log.debug(
+                    f"\n\nORIGINAL:\n{self.df[col].value_counts()}\n\nENCODED:\n{self.df[name].value_counts()}\n"
+                )
         self.rejoin_original()
         return self.df