New functions and updated readme

radwinskis · Jan 19, 2022 · 906ff3a · 906ff3a
1 parent 2fcf7ed
commit 906ff3a
Show file tree

Hide file tree

Showing 4 changed files with 122 additions and 4 deletions.
diff --git a/README.md b/README.md
@@ -25,11 +25,13 @@ One can install RSAbs using pip or by copying the GitHub repo to their working d
 pip install RSAbs
 ```
 ## Function Parameters
-There are two functions which make up the RSAbs package. 
+There are four functions which make up the RSAbs package. 
 1) ```sh
     absorptions(spectra_df, wavelength_df, integrity='ASD_CR') 
     ``` 
-spectra_df: dataframe with spectra as columns
+Functionality: creates DataFrame of all absorption wavelength positions
+
+spectra_df: dataframe with spectra as columns (leave out wavelength column, if applicable)
 
 wavelength_df: dataframe with wavelength as column
 
@@ -40,6 +42,9 @@ If you desire compatibility with another sensor/data type, please let me know.
 2) ```sh
     seek_absorptions(absorptions_df, start_wavelength, end_wavelength, wavelength_unit='nanometer') 
     ```
+
+Functionality: creates narrowed down DataFrame of absorption wavelength positions in a given wavelength range, and only returns data for samples that contain an absorption in the given range.
+
 absorptions_df: dataframe from absorptions() function
 
 start_wavelength: starting wavelength value for range of interest (smaller wavelength)
@@ -48,6 +53,36 @@ end_wavelength: ending wavelength value for range of interest (larger wavelength
 
 wavelength_unit: specify unit of wavelength. Options: 'nanometer' OR 'micrometer'. PLEASE ensure these are chosen correctly. Default is 'nanometer'.
 
+3) ```sh
+    absorptions_depth(spectra_df, integrity='ASD_CR') 
+    ```
+
+Functionality: creates DataFrame of all absorption depths
+
+spectra_df: dataframe with spectra as columns (leave out wavelength column, if applicable)
+
+integrity: spectra type, indicates resolution and continuum removal status. Options: 'ASD_CR' for continuum removed ASD data, 'ASD_normal' for non-continuum-removed ASD data, or 'HyMap_CR' for continuum-removed HyMap data. Default is 'ASD_CR'
+
+4) ```sh
+    seek_absorptions_depth(spectra_df, wavelength_df, integrity='ASD_CR', start_wavelength=350, end_wavelength=2500, wavelength_unit='nanometer') 
+    ```
+
+Functionality: creates narrowed down DataFrame of absorption depths in a given wavelength range,
+and only returns data for samples that contain an absorption in the given wavelength range. 
+
+spectra_df: dataframe with spectra as columns (leave out wavelength column, if applicable)
+
+wavelength_df: dataframe with wavelength as column
+
+integrity: spectra type, indicates resolution and continuum removal status. Options: 'ASD_CR' for continuum removed ASD data, 'ASD_normal' for non-continuum-removed ASD data, or 'HyMap_CR' for continuum-removed HyMap data. Default is 'ASD_CR'
+
+start_wavelength: starting wavelength value for range of interest (smaller wavelength)
+
+end_wavelength: ending wavelength value for range of interest (larger wavelength)
+
+wavelength_unit: specify unit of wavelength. Options: 'nanometer' OR 'micrometer'. PLEASE ensure these are chosen correctly. Default is 'nanometer'.
+
+
 ## Usage 
 
 To use the functions, either import the RSAbs package and use 

diff --git a/RSAbs/absorptions_depth.py b/RSAbs/absorptions_depth.py
@@ -0,0 +1,23 @@
+from scipy.signal import find_peaks
+import pandas as pd
+import numpy as np
+import itertools
+
+def absorptions_depth(spectra_df, integrity='ASD_CR'): #returns dataframe with the depth of each feature, where the depth is from the top of the convex hull (1.00) minus the value of the absorption minima
+    if isinstance(spectra_df, pd.DataFrame)==False: #Check whether spectra_df argument is satisfactory
+            print('Error: spectra_df argument not identified as pd.DataFrame')
+    def depth_props(spectra_df):        
+        spectra_array = spectra_df.values.flatten()
+        if integrity=='ASD_CR':
+            absorps, props = find_peaks(-spectra_array, width = 6, distance = 25, prominence = 0.001) #For using continuum removed spectra
+        elif integrity=='ASD_Normal':
+            absorps, props = find_peaks(-spectra_array, width = 5, distance = 15, prominence = 0.001) #For using non-continuum removed spectra
+        elif integrity=='HyMap_CR':
+            absorps, props = find_peaks(-spectra_array, width = 3, distance = 5, prominence = 0.001) #For using continuum removed HYMAP spectra
+        else:
+            print('Enter integrity argument based on avaiable options:"ASD_CR" for ASD continuum removed spectra, "ASD_Normal" for ASD non-continuum removed spectra, or "HyMap_CR" for HyMap continuum removed spectra. Default is "ASD_CR".')
+
+        depths = props["prominences"]
+        return pd.Series(depths) #For normal spectra
+
+    return pd.DataFrame(spectra_df.apply(depth_props, axis=0))
diff --git a/RSAbs/seek_absorptions_depth.py b/RSAbs/seek_absorptions_depth.py
@@ -0,0 +1,60 @@
+from scipy.signal import find_peaks
+import pandas as pd
+import numpy as np
+import itertools
+
+def seek_absorptions_depth(spectra_df, wavelength_df, integrity='ASD_CR', start_wavelength=350, end_wavelength=2500, wavelength_unit='nanometer'):
+    if isinstance(spectra_df, pd.DataFrame)==False: #Check whether spectra_df argument is satisfactory
+            print('Error: spectra_df argument not identified as pd.DataFrame')
+    if isinstance(wavelength_df, pd.DataFrame)==False: #Check whether spectra_df argument is satisfactory
+            print('Error: spectra_df argument not identified as pd.DataFrame')
+    def abs_features(spectra_df, wavelength_df):        
+        spectra_array = spectra_df.values.flatten()
+        wavelength_array = wavelength_df.values.flatten()
+        if integrity=='ASD_CR':
+            absorps, props = find_peaks(-spectra_array, width = 6, distance = 25, prominence = 0.001) #For using continuum removed spectra
+        elif integrity=='ASD_Normal':
+            absorps, props = find_peaks(-spectra_array, width = 5, distance = 15, prominence = 0.001) #For using non-continuum removed spectra
+        elif integrity=='HyMap_CR':
+            absorps, props = find_peaks(-spectra_array, width = 3, distance = 5, prominence = 0.001) #For using continuum removed HYMAP spectra
+        else:
+            print('Enter integrity argument based on avaiable options:"ASD_CR" for ASD continuum removed spectra, "ASD_Normal" for ASD non-continuum removed spectra, or "HyMap_CR" for HyMap continuum removed spectra. Default is "ASD_CR".')
+
+        absorp_locs = wavelength_array[absorps]
+        depths = props["prominences"]
+        return pd.Series(absorp_locs) #For normal spectra
+
+    abs_wvls = pd.DataFrame(spectra_df.apply(abs_features, wavelength_df=wavelength_df, axis=0))
+
+    def abs_depths(spectra_df, wavelength_df):        
+        spectra_array = spectra_df.values.flatten()
+        wavelength_array = wavelength_df.values.flatten()
+        if integrity=='ASD_CR':
+            absorps, props = find_peaks(-spectra_array, width = 6, distance = 25, prominence = 0.001) #For using continuum removed spectra
+        elif integrity=='ASD_Normal':
+            absorps, props = find_peaks(-spectra_array, width = 5, distance = 15, prominence = 0.001) #For using non-continuum removed spectra
+        elif integrity=='HyMap_CR':
+            absorps, props = find_peaks(-spectra_array, width = 3, distance = 5, prominence = 0.001) #For using continuum removed HYMAP spectra
+        else:
+            print('Enter integrity argument based on avaiable options:"ASD_CR" for ASD continuum removed spectra, "ASD_Normal" for ASD non-continuum removed spectra, or "HyMap_CR" for HyMap continuum removed spectra. Default is "ASD_CR".')
+
+        absorp_locs = wavelength_array[absorps]
+        depths = props["prominences"]
+        return pd.Series(depths) #For normal spectra
+
+    abs_depths = pd.DataFrame(spectra_df.apply(abs_depths, wavelength_df=wavelength_df, axis=0))
+
+    wvls_range = end_wavelength-start_wavelength
+    if wavelength_unit=='nanometer':
+        steps = int(wvls_range+1)
+    elif wavelength_unit=='micrometer':
+        steps = int(wvls_range*1000+1)
+    else: 
+        print('Error: Incorrect wavelength_unit argument. Choose from "nanometer" or "micrometer" and ensure they are the correct units. Default is set to nanometer.')
+    wvls_array = np.linspace(start_wavelength, end_wavelength, num=steps)
+    seek_samples = abs_wvls.columns[abs_wvls.isin(wvls_array).any()]
+    samples = abs_wvls[seek_samples]
+    wvls_of_samples_of_interest = samples[samples[seek_samples].isin(wvls_array)].apply(lambda x: pd.Series(x.dropna().values))
+    depths_of_samples_of_interest = abs_depths[seek_samples]
+    depths_final = depths_of_samples_of_interest[samples[seek_samples].isin(wvls_array)].apply(lambda x: pd.Series(x.dropna().values))
+    return depths_final
diff --git a/setup.py b/setup.py
@@ -2,13 +2,13 @@
 setup(
   name = 'RSAbs',       
   packages = ['RSAbs'],   
-  version = '0.7', 
+  version = '0.8', 
   license='MIT',  
   description = 'AUTOMATIC DETECTION OF ABSORPTION FEATURES IN REFLECTANCE SPECTRA', 
   author = 'MARK RADWIN',  
   author_email = 'markradwin@gmail.com',     
   url = 'https://github.com/radwinskis/RSAbs',  
-  download_url = 'https://github.com/radwinskis/RSAbs/archive/refs/tags/v0.7.tar.gz',   
+  download_url = 'https://github.com/radwinskis/RSAbs/archive/refs/tags/v0.8.tar.gz',   
   keywords = ['REFLECTANCE', 'SPECTROSCOPY', 'ABSORPTION', 'DETECTION'],  
   install_requires=[          
           'scipy',