Let uploadresults.py execute computerankings.py itself as appropriate.

2 files changed, 112 insertions, 19 deletions

diff --git a/scripts/finalizeresults.py b/scripts/computerankings.py
index dafa858..6726806 100755
--- a/scripts/finalizeresults.py
+++ b/scripts/computerankings.py
@@ -12,8 +12,8 @@ from misc import (
 def printUsage():
     print (
         "usage:" + sys.argv[0] +
-        " --help | [--dbhost H --dbport P] --db D --host H --platform P "
-        "--branch B --sha1 S")
+        " --help | [--dbhost H] [--dbport P] --db D --host H --platform P "
+        "--branch B --sha1 S [--noprogress NP]")
 
 def printVerboseUsage():
     printUsage()
@@ -40,6 +40,8 @@ def printVerboseUsage():
         "    --sha1: The tested revision within the branch. Can be "
         "extracted using 'git log -1 --pretty=format:%H' (assuming the "
         "tested revision is the current head revision).")
+    print(
+        "  --noprogress: Specify \'true\' to disable progress indicator.")
 
 
 # ### 2 B DOCUMENTED!
@@ -188,7 +190,8 @@ def getFirstUploadTimestamp(snapshots, sha1_id):
 
 
 # ### 2 B DOCUMENTED!
-def updateRankings(host_id, platform_id, branch_id, sha12_id, context2_id):
+def updateRankings(
+    host_id, platform_id, branch_id, sha12_id, context2_id, no_progress):
 
     # Get all snapshots matching the host/platform/branch combination:
     sys.stdout.write("getting snapshots ... ")
@@ -263,11 +266,17 @@ def updateRankings(host_id, platform_id, branch_id, sha12_id, context2_id):
         sys.exit(1)
 
     # Get time series statistics for all benchmarks:
+    if no_progress:
+        sys.stdout.write("getting time series statistics ... ")
     bmstats_list = getBMTimeSeriesStatsList(
         host_id, platform_id, branch_id, snapshots, None, difftol, durtolmin,
-        durtolmax, printProgress, "getting time series statistics")
+        durtolmax, None if no_progress else printProgress,
+        "getting time series statistics")
 
-    sys.stdout.write("\n")
+    if no_progress:
+        sys.stdout.write("done\n")
+    else:
+        sys.stdout.write("\n")
 
 
     # *** Compute rankings **************************************************
@@ -288,12 +297,12 @@ def updateRankings(host_id, platform_id, branch_id, sha12_id, context2_id):
         sys.exit(1)
     nameToIndex = { "QS": 3, "LCSSR": 4, "LCSSI": 5, "LCSS1R": 6, "LCSS1I": 7 }
     for name in nameToIndex:
-        sys.stdout.write("registering ranking for " + name + " ...\r")
+        sys.stdout.write("registering ranking for " + name + " ... ")
         sys.stdout.flush()
         registerRanking(
             table, nameToIndex[name], name, context1_id, context2_id)
-
-    sys.stdout.write("\n")
+        sys.stdout.write("done\n")
+        sys.stdout.flush()
 
 # --- END Global functions ----------------------------------------------
 
@@ -342,12 +351,15 @@ if context2_id == -1:
     print "error: no results found for this context"
     sys.exit(1)
 
-
-updateRankings(host_id, platform_id, branch_id, sha12_id, context2_id)
+updateRankings(
+    host_id, platform_id, branch_id, sha12_id, context2_id,
+    ("noprogress" in options) and (
+        (options["noprogress"] == "1")
+        or (options["noprogress"].lower() == "true")))
 
 # Write to database:
 commit()
 
-print "finalization done"
+print "rankings computation done"
 
 # --- END Main program ----------------------------------------------
diff --git a/scripts/uploadresults.py b/scripts/uploadresults.py
index 3c92691..f0db5a7 100755
--- a/scripts/uploadresults.py
+++ b/scripts/uploadresults.py
@@ -1,9 +1,11 @@
 #!/usr/bin/env python
 
 import sys
+from subprocess import Popen, PIPE
 from xml.dom.minidom import parse
 from dbaccess import setDatabase, execQuery, commit
-from misc import getOptions, textToId, isValidSHA1, getContext
+from misc import (
+    getOptions, textToId, idToText, isValidSHA1, getContext, getAllSnapshots)
 
 
 # --- BEGIN Global functions ----------------------------------------------
@@ -253,6 +255,75 @@ def rankingsExist(host, platform, branch, sha1):
     return len(matches) > 0
 
 
+# Returns the context ID if found, otherwise -1:
+def getContextIdFromNames(options):
+    host_id = textToId("host", options["host"])
+    platform_id = textToId("platform", options["platform"])
+    branch_id = textToId("branch", options["branch"])
+    sha1_id = textToId("sha1", options["sha1"])
+    return getContext(host_id, platform_id, branch_id, sha1_id)
+
+
+# Returns True iff this context exists:
+def contextExists(options):
+    return getContextIdFromNames(options) != -1
+
+
+# Returns True iff no more results are to be expected for this context:
+def contextComplete(options):
+
+    max_sample_size = 5 # WARNING: This value must match the corresponding value
+                        # in the script that triggers benchmark execution.
+
+    context_id = getContextIdFromNames(options)
+    sample_size = execQuery(
+        "SELECT count(*) FROM"
+        " (SELECT DISTINCT uploadId from result where contextId=%d) AS foo;"
+        % context_id)[0][0]
+
+    return sample_size >= max_sample_size
+
+
+# Executes the external computerankings.py script with appropriate arguments.
+def execComputeRankings(options, use_latest_snapshot):
+
+    if use_latest_snapshot:
+        # Attempt to use the latest available snapshot for this
+        # host/platform/branch combination as the actual snapshot:
+        host_id = textToId("host", options["host"])
+        platform_id = textToId("platform", options["platform"])
+        branch_id = textToId("branch", options["branch"])
+        snapshots = getAllSnapshots(host_id, platform_id, branch_id)
+        if len(snapshots) > 0:
+            actual_snapshot = idToText("sha1", snapshots[-1][0])
+        else:
+            return # special case when no snapshots exist yet
+    else:
+        actual_snapshot = options["sha1"]
+
+    cmd = [
+        "computerankings.py", "--db", options["db"], "--host", options["host"],
+        "--platform", options["platform"], "--branch", options["branch"],
+        "--sha1", actual_snapshot, "--noprogress", "true"]
+    if "dbhost" in options:
+        cmd += ["--dbhost", options["dbhost"]]
+    if "dbport" in options:
+        cmd += ["--dbport", options["dbport"]]
+
+    p = Popen(cmd, stdout = PIPE, stderr = PIPE)
+    stdout, stderr = p.communicate()
+    if (p.returncode != 0):
+        print "failed to execute command '" + cmd + "':"
+        print "  return code:", p.returncode
+        print "  stdout: >%s<" % stdout.strip()
+        print "  stderr: >%s<" % stderr.strip()
+    else:
+        print "computerankings.py executed successfully:"
+        print "  return code:", p.returncode
+        print "  stdout: >%s<" % stdout.strip()
+        print "  stderr: >%s<" % stderr.strip()
+
+
 # --- END Global functions ----------------------------------------------
 
 
@@ -281,27 +352,37 @@ setDatabase(
 # Reject uploading if rankings exist for this context, since that would
 # require a recomputation of these rankings.
 #
-# (Note that this check if sufficient as long as we assume that results
-# are uploaded in an order that is consistent with the order of the
-# corresponding SHA-1s in the branch in question (i.e. that results will
-# never be uploaded for a SHA-1 that precedes (is "older" than) a SHA-1 for
-# which results already exist). This assumption ensures that results for a
-# new SHA-1 will not alter the input for any existing ranking.)
+# (Note that we only need to check the current SHA-1 as long as we assume that
+# results are uploaded in an order that is consistent with the order of the
+# corresponding SHA-1s in the branch in question: Rankings for earlier SHA-1s
+# will not be affected by results for this SHA-1, and rankings for later
+# SHA-1s cannot exist a this point (by assumption).)
 if rankingsExist(
     options["host"], options["platform"], options["branch"], options["sha1"]):
     print "error: this context has already been finalized"
     sys.exit(1)
 
+# If this is the first set of results for the current context, we finalize the
+# results for the same context, only one snapshot back:
+if not contextExists(options):
+    execComputeRankings(options, True)
+
+# Parse results and store them in the database:
 try:
     results = extractResults(options["file"])
 except BaseException as e:
     print "error: failed to parse XML file:", e.args[0]
     sys.exit(1)
-
 uploadToDatabase(
     options["host"], options["platform"], options["branch"], options["sha1"],
     results)
 
 print "uploading done"
 
+# If no more results are expected in this context, we can finalize the
+# results already at this point:
+if contextComplete(options):
+    execComputeRankings(options, False)
+
+
 # --- END Main program ----------------------------------------------