Merge branch 'master' of https://github.com/wjlei1990/pypaw

Author: wjlei1990

scripts/window_weights/README.md

@@ -1,24 +1,31 @@
-### Introduction
+## Introduction
 
 This directory contains example files for two different weighting strategies,
 both of which are already implemented in pypaw.
 
-##### Weighting Strategy Version I
+---
+#### Weighting Strategy Version I
 
 The weightings is determined in order of category, receiver and finally, source.
   1. determine the category weightings ratio based on number of window counts in each category.
+  
     Attention here that we determined is the ratio between each category but not the absolute
     values. The absolut values will be determined on each event level(to satisfy the
     normalization equation).
+    
   2. determine the receiver weightings based on station geographic distribution(ratios) and window
     counts(normalization) of each receiver given one event.
+    
     Also, determine the absolute values of category weightings for each source.
     Once you reach this step, you can use the weights
     (category and receiver combined) to sum the adjoint source from each category for each event.
     And once you get summed adjoint source for each event, you can then launch the adjoint simulation
     to calculate the kernel for each source.
+    
   3. determine the source weightings based on source geographic distribution(ratio) and
-    number of windows(normalization) of each source. The source weightings are applied on kernels
+    number of windows(normalization) of each source.
+    
+    The source weightings are applied on kernels
     when summing all the kernels together.
 
 The benefit of version I is that you can freely combine kernels from different number of sources.
@@ -34,16 +41,22 @@ this weighting gives you more freedom when combining kernels.
 The draw back of that is it doesn't gurantee the weightins for different categories are
 evenly balanced(but will not be too deviated, based on my experiment).
 
+---
 
-###### Weighting Strategy Version II
+#### Weighting Strategy Version II
 The weightings are determined in order of receiver, source and finally category.
   1. determine the receiver weightings based on window counts and station distribution.
     (same as version I)
+    
   2. determine the source weightings based on source geographic distribution(ratio) and
     number of sources(the sum of source weights eqaul to number of sources)
+    
   3. determine the category weightings based on source weights(ws) number of windows in
     each category(Nsc) and number of categories(C).
-      wc = 1 / (C * \sum_{s} ws * Nsc)
+    
+    ```
+      W_{} = 1 / (C * \sum_{s} ws * N_{sc})
+    ```
   4. Finally, you combine the receiver, source and category weightings
 
 The benefit of this weighting is you can guarantee that the sum of weightings from each

scripts/window_weights/version_I/01_category_weights/README.md

@@ -9,7 +9,7 @@
   Then, run the script to generate path file.
 
   ```
-    python generate_count_windows.py -f cmtlist.txt
+    python generate_count_windows.py
   ```
 
 2. Run the script: