From d74ed3f31f1cd29c965db540ff7a53adc55e02f3 Mon Sep 17 00:00:00 2001
From: "R. Biswas" <rbiswas4@gmail.com>
Date: Wed, 6 Apr 2016 18:08:33 -0700
Subject: [PATCH] Adding module that produces a subset of the data repository
 containing only a subset of visits.

	new file:   copyData.py
---
 bin/copyData.py | 82 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 82 insertions(+)
 create mode 100644 bin/copyData.py

diff --git a/bin/copyData.py b/bin/copyData.py
new file mode 100644
index 0000000..ea1cfbc
--- /dev/null
+++ b/bin/copyData.py
@@ -0,0 +1,82 @@
+"""
+Copies parts of a data repository corresponding to certain visits into a new
+repository to make a small test repository
+
+To Use change the strings in script part:
+    - change the srcdir to point to the repo that is to be copied
+    - obtain the list visits to a list of obshistIDs of visits that are desired in the subset
+    - change newRepo to point to the absolute path of location where you want this 
+"""
+from __future__ import absolute_import
+import shutil
+import os
+import glob
+import pandas as pd
+
+
+def copyRepoForVisits(srcRepo, newRepo, visitSeq, dirDepth=3):
+    """
+    Copy a part of source repository srcRepo to a new Repo
+    keeping only information corresponding to visits iDs in
+    visitSeq.
+    This is done by using the structure of the repository
+    and not by directory structure independent methods. It copies
+    directories and files which have the format 'vvisitID' within
+    a depth of 3 in the source repository.
+
+    Parameters
+    ----------
+    srcRepo : str, mandatory
+        absolute path to data repository from which results are to be
+        copied
+    newRepo : str, mandatory
+        absolute path to location where newrepository is to be written
+    visitSeq : Iterable of integers, mandatory
+        sequence of integers indexing the visits that should be copied
+    dirDepth : integer, optional, defaults to 3
+        Depth to which the directory of the form v visitID will be
+        searched for
+
+    Returns
+    -------
+    list of integers indexing visits that were requested but not found
+    in the original data repository
+
+    Examples
+    --------
+    """
+    if not os.path.exists(newRepo):
+        os.makedirs(newRepo)
+    # list of directory trees to copy over
+    srcs = []
+    # list of visits that have not been found
+    notFoundList = []
+    newRepo = os.path.abspath(newRepo)
+
+    for visit in visitSeq:
+        visitFound = False
+        for depth in range(dirDepth):
+            mystr = srcRepo + '/*'*depth + '/v' + visit + '*'
+            src = glob.glob(mystr)
+            if len(src) > 0:
+                srcs += src
+                visitFound = True
+        if not visitFound:
+            notFoundList.append(visit)
+    for src in srcs:
+        dest = os.path.join(newRepo, src.split(srcdir)[1][1:])
+        shutil.copytree(src, dest)
+    return notFoundList
+
+if __name__ == '__main__':
+
+    # Generate the visits (obsHistIDs)
+    df = pd.read_csv('obsHistIDs.csv').head(5)
+    visits = [str(id) for id in df.obsHistID.values]
+
+    # Directories
+    srcdir = '/global/homes/t/tony_j/Twinkles/985visits/985visits'
+    newRepo = 'repo4'
+
+    vnf = copyRepoForVisits(srcRepo=srcdir, newRepo=newRepo, visitSeq=visits)
+    print(vnf)