Add basic dxt support (without segments). Updates to documentation, and example notebooks.

darshan-util/pydarshan/Makefile

@@ -43,6 +43,8 @@ coverage: ## check code coverage quickly with the default Python
 
 docs: ## generate Sphinx HTML documentation, including API docs
 	rm -f docs/darshan.rst
+	rm -f docs/darshan.backend.rst
+	rm -f docs/darshan.plots.rst
 	rm -f docs/modules.rst
 	sphinx-apidoc -o docs/ darshan
 	$(MAKE) -C docs clean

darshan-util/pydarshan/darshan/__init__.py

@@ -5,7 +5,5 @@
 __version__ = '0.1.0'
 
 
-from darshan.parser_cffi import *
-
-
+from darshan.backend.cffi_backend import *
 from darshan.report import *

darshan-util/pydarshan/darshan/api_def_c.py

@@ -100,6 +100,19 @@ struct darshan_apxc_perf_record
 
 
 
+struct dxt_file_record {
+    struct darshan_base_record base_rec;
+    int64_t shared_record;  /* -1 means it is a shared file record */
+    char hostname[64];      /* size defined via macro */
+
+    int64_t write_count;
+    int64_t read_count;
+};
+
+
+
+
+
 typedef uint64_t darshan_record_id;
 
 struct darshan_name_record

darshan-util/pydarshan/darshan/parser_ascii.py → darshan-util/pydarshan/darshan/backend/ascii_backend.py

@@ -642,7 +642,7 @@ class AsciiDarshanLog(object):
         logdata = self.logdata
 
 
-        # Original:
+        # Original, Target:
         ## <type>, <app time>, <read>, <write>, <meta>
         #POSIX, 98.837925, 0.150075, 0.5991, 0.4129
         #MPI-IO, 97.293875, 0.051575, 0.126525, 2.528025

darshan-util/pydarshan/darshan/parser_cffi.py → darshan-util/pydarshan/darshan/backend/cffi_backend.py

@@ -158,6 +158,57 @@ def log_get_name_records(log):
 
 
 
+def log_get_dxt_record(log, mod_name, mod_type, mode='numpy'):
+    """
+    Returns a dictionary holding a dxt darshan log record.
+
+    Args:
+        log: Handle returned by darshan.open
+        mod_name (str): Name of the Darshan module
+        mod_type (str): String containing the C type
+
+    Return:
+        dict: generic log record
+
+    Example
+    -------
+
+    The typical darshan log record provides two arrays, on for integer counters
+    and one for floating point counters:
+
+    >>> darshan.log_get_dxt_record(log, "POSIX", "struct darshan_posix_file **")
+    {'rank': 42, 'read_segments': array([...])}
+
+
+    """
+
+
+
+    modules = log_get_modules(log)
+
+    rec = {}
+    buf = ffi.new("void **")
+    r = libdutil.darshan_log_get_record(log, modules[mod_name]['idx'], buf)
+    if r < 1:
+        return None
+    rbuf = ffi.cast(mod_type, buf)
+    clst = []
+
+    rec['id'] = rbuf[0].base_rec.id
+    rec['rank'] = rbuf[0].base_rec.rank
+    rec['write_count'] = rbuf[0].write_count
+    rec['read_count'] = rbuf[0].read_count
+    
+    rec['write_segments'] = []
+    rec['read_segments'] = []
+
+    return rec
+
+
+
+
+
+
 def log_get_generic_record(log, mod_name, mod_type, mode='numpy'):
     """
     Returns a dictionary holding a generic darshan log record.
@@ -320,6 +371,7 @@ def log_get_posix_record(log):
     return log_get_generic_record(log, "POSIX", "struct darshan_posix_file **")
 
 
+
 def log_get_stdio_record(log):
     """
     Returns a darshan log record for STDIO.

darshan-util/pydarshan/darshan/error.py

+""" Darshan Error classes and functions. """
+
+
+
+class DarshanBaseError(Exception):
+    """
+    Base exception class for Darshan errors in Python.
+    """
+    pass
+
+
+class DarshanVersionError(NotImplementedError):
+    """
+    Raised when using a features is used which is not provided by libdarshanutil.
+    """
+    min_version = None
+
+    def __init__(self, min_version, msg="Feature"):
+        self.msg = msg
+        self.min_version = min_version
+        self.version = "0.0.0"
+
+    def __repr__(self):
+        return "DarshanVersionError('%s')" % str(sefl)
+
+    def __str__(self):
+        return "%s requires libdarshanutil >= %s, have %s" % (self.msg, self.min_version, self.version)
+

darshan-util/pydarshan/darshan/report.py

 #!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 
-import darshan.parser_cffi as parser
+import darshan.backend.cffi_backend as backend
 import json
 import numpy as np
 import re
+import copy
+
 
 class NumpyEncoder(json.JSONEncoder):
     def default(self, obj):
@@ -15,43 +17,147 @@ class NumpyEncoder(json.JSONEncoder):
 
 
 class DarshanReport(object):
+    """
+    The DarshanReport class provides a convienient wrapper to access darshan
+    logs, which also caches already fetched information. In addition to that
+    a number of common aggregations can be performed.
+    """
 
     def __init__(self, filename):
         self.filename = filename
 
         self.report = {'version': 1}
+        self.report['records'] = {}
+
+
         self.filehashes = {}
 
-        self.log = parser.log_open(self.filename)
+        self.log = backend.log_open(self.filename)
+
+        # state dependent book-keeping
+        self.converted_records = False  # true if convert_records() was called (unnumpyfiy)
+
+        # when using report algebra this log allows to untangle potentially
+        # unfair aggregations (e.g., double accounting)
+        self.provenance_log = []
+
+
+
+        self.read_metadata()
+        self.report["name_records"] = backend.log_get_name_records(self.log)
+
+
+
+    def __add__(self, other):
+
+        new_report = self.copy()
+        #new_report = copy.deepcopy(self)
+        new_report.provenance_log.append(("add", self, other))
+
+
+        return new_report
+
 
 
     def read_all(self):
         """
         Read all available information from darshan log and return das dictionary.
-        """
-
-        self.report["job"] = parser.log_get_job(self.log)
-        self.report["exe"] = parser.log_get_exe(self.log)
-        self.report["mounts"] = parser.log_get_mounts(self.log)
-        self.report["modules"] = parser.log_get_modules(self.log)
 
-        self.report["name_records"] = parser.log_get_name_records(self.log)
+        Args:
+            None
 
+        Return:
+            None
+        """
 
-        self.report['records'] = {}
         for mod in self.report['modules']:
             self.mod_read_all_records(mod)
 
+        pass
+
+
+    def read_metadata(self):
+        """
+        Read metadata such as the job, the executables and available modules.
+
+        Args:
+            None
+
+        Return:
+            None
+
+        """
+        self.report["job"] = backend.log_get_job(self.log)
+        self.report["exe"] = backend.log_get_exe(self.log)
+        self.report["mounts"] = backend.log_get_mounts(self.log)
+        self.report["modules"] = backend.log_get_modules(self.log)
+
+
+
+    def mod_read_all_dxt_records(self, mod, mode='numpy'):
+        """
+        Reads all records for module
+
+        Args:
+            mod (str): Identifier of module to fetch all records
+            mode (str): 'numpy' for ndarray (default), 'dict' for python dictionary
+
+        Return:
+            None
+
+        """
+        supported =  ['DXT_POSIX', 'DXT_MPIIO']
+
+        if mod not in supported:
+            print("Skipping. Currently unsupported:", mod)
+            # skip mod
+            return 
+
+
+        structdefs = {
+            "DXT_POSIX": "struct dxt_file_record **",
+            "DXT_MPIIO": "struct dxt_file_record **",
+        }
+
+
+        self.report['records'][mod] = []
+        self.report['modules'][mod]['num_records'] = 0
+
+
+
 
-        self.summarize()
 
 
-        self.convert_records()
+        rec = backend.log_get_dxt_record(self.log, mod, structdefs[mod])
+        while rec != None:
+            # TODO: performance hog and hacky ;)
+            #recs = json.dumps(rec, cls=NumpyEncoder)
+            #rec = json.loads(recs)
+
+            if mode == 'numpy': 
+                self.report['records'][mod].append(rec)
+            else:
+                print("Not implemented.")
+                exit(1)
+
+                #c = dict(zip(cn, rec['counters']))
+                #fc = dict(zip(fcn, rec['fcounters']))
+                #self.report['records'][mod].append([c, fc])
+                pass
+
+
+            self.report['modules'][mod]['num_records'] += 1
+
+            # fetch next
+            rec = backend.log_get_dxt_record(self.log, mod, structdefs[mod])
 
         pass
 
 
 
+
+
+
     def mod_read_all_records(self, mod, mode='numpy'):
         """
         Reads all records for module
@@ -64,12 +170,14 @@ class DarshanReport(object):
             None
 
         """
+        unsupported =  ['DXT_POSIX', 'DXT_MPIIO', 'LUSTRE']
 
-        unsupported =  ['DXT_POSIX', 'DXT_MPIIO']
         if mod in unsupported:
+            print("Skipping. Currently unsupported:", mod)
             # skip mod
             return 
 
+
         structdefs = {
             "BG/Q": "struct darshan_bgq_record **",
             "HDF5": "struct darshan_hdf5_file **",
@@ -77,20 +185,26 @@ class DarshanReport(object):
             "PNETCDF": "struct darshan_pnetcdf_file **",
             "POSIX": "struct darshan_posix_file **",
             "STDIO": "struct darshan_stdio_file **",
-            "DECAF": "struct darshan_decaf_record **"
+            "DECAF": "struct darshan_decaf_record **",
+            "DXT_POSIX": "struct dxt_file_record **",
         }
 
 
         self.report['records'][mod] = []
 
-        cn = parser.counter_names(mod)
-        fcn = parser.fcounter_names(mod)
+        cn = backend.counter_names(mod)
+        fcn = backend.fcounter_names(mod)
 
         self.report['modules'][mod]['counters'] = cn 
         self.report['modules'][mod]['fcounters'] = fcn
         self.report['modules'][mod]['num_records'] = 0
 
-        rec = parser.log_get_generic_record(self.log, mod, structdefs[mod])
+
+
+
+
+
+        rec = backend.log_get_generic_record(self.log, mod, structdefs[mod])
         while rec != None:
             # TODO: performance hog and hacky ;)
             #recs = json.dumps(rec, cls=NumpyEncoder)
@@ -107,12 +221,44 @@ class DarshanReport(object):
             self.report['modules'][mod]['num_records'] += 1
 
             # fetch next
-            rec = parser.log_get_generic_record(self.log, mod, structdefs[mod])
+            rec = backend.log_get_generic_record(self.log, mod, structdefs[mod])
 
         pass
 
 
 
+    def mod_agg(self, mod, ranks=None, files=None, preserve_rank=False, preserve_file=False):
+        """
+        Aggregate counters for a given module name and return updated dictionary.
+        
+        Args:
+            mod (str): Name of the mod to aggregate.
+            ranks (int or list): Only aggregate if rank is matched
+            files (int or list): Only aggregate if file is matched
+            preserve_rank: do not collapse ranks into single value
+            preserve_file: do not collapse files into single value
+
+        Return:
+            List of aggregated records
+        """
+
+
+        # TODO: assert
+
+        c = None
+        fc = None
+
+        # aggragate
+        for rec in recs[mod]:
+            if mod not in ctx:
+                c = rec['counters']
+                fc = rec['counters']
+            else:
+                c = np.add(ctx[mod], rec['counters'])
+                fc = np.add(ctx[mod], rec['fcounters'])
+
+        return {'counters': c, 'fcounter': fc}
+
 
     def mod_agg_iohist(self, mod, mode='append'):
         """
@@ -125,15 +271,19 @@ class DarshanReport(object):
             None
         """
 
+        # convienience
+        recs = self.report['records']
+        ctx = {}
+
+
         supported = ["POSIX", "MPI-IO"]
         if mod not in supported:
             raise Exception("Unsupported mod_name for aggregated iohist.")
 
 
-        # convienience
-        recs = self.report['records']
-        ctx = {}
-
+        # check records for module are present
+        if mod not in recs:
+            return
 
 
         # aggragate
@@ -158,11 +308,11 @@ class DarshanReport(object):
 
         tmp = json.dumps(ctx[mod], cls=NumpyEncoder)
         tmp = json.loads(tmp)
-        cn = parser.counter_names(mod)
+        cn = backend.counter_names(mod)
         c = dict(zip(cn, tmp))
         c = {k:c[k] for k in c if re.match('.*?_SIZE_.*?', k)}
         c = {fix_name(k):v for k, v  in c.items()}
-        ctx[mod] = c
+        ctx = c
 
 
         if mode == 'append':
@@ -200,6 +350,11 @@ class DarshanReport(object):
         # aggragate
         mods = ['MPI-IO', 'POSIX', 'STDIO']
         for mod in mods:
+
+            # check records for module are present
+            if mod not in recs:
+                continue
+
             agg = None
             for rec in recs[mod]:
                 if agg is not None:
@@ -209,7 +364,7 @@ class DarshanReport(object):
 
 
             # filter fields
-            cn = parser.counter_names(mod)
+            cn = backend.counter_names(mod)
             agg = dict(zip(cn, agg.tolist()))
             
 
@@ -277,11 +432,150 @@ class DarshanReport(object):
             return ctx
 
 
+
+
+    def create_sankey(self):
+        """
+        Generate a summary that shows the dataflow from ranks to files to mnts.
+        """
+
+        # convienience
+        recs = self.report['records']
+        nrecs = self.report['name_records']
+        ctx = {}
+
+
+        # check records for module are present
+        if 'POSIX' not in recs:
+            self.report['sankey'] = None
+            return
+
+        ranks = {}
+        files = {}
+        mounts = {}
+
+        nodes = {}
+        edges = {}
+
+
+        # build mnt list
+        mnts = []
+        for mnt in self.report['mounts']:
+            mnts.append(mnt[0])
+
+
+        # collect records
+        for rec in recs['POSIX']:
+            
+            rnk = "r_%d" % (rec['rank'])
+
+            fnr = rec['id']
+            fnr = nrecs[fnr]
+
+            # determine mount point
+            mnt = None
+            for curr in mnts:
+                if re.search(curr, fnr):
+                    mnt = curr
+                    break
+            mnt = "m_%s" % mnt
+
+           
+            nodes[rnk] = {'name': rnk}
+            nodes[fnr] = {'name': fnr}
+            nodes[mnt] = {'name': mnt}
+
+            #rnk2fnr += 1
+            #fnr2mnt += 1
+
+
+            rnk2fnr = "%s->%s" % (rnk, fnr)
+            fnr2mnt = "%s->%s" % (fnr, mnt)
+
+            if rnk2fnr not in edges:
+                edges[rnk2fnr] = {"value": 0, "source": rnk, "target": fnr}
+            edges[rnk2fnr]["value"] += 1
+
+
+            if fnr2mnt not in edges:
+                edges[fnr2mnt] = {"value": 0, "source": fnr, "target": mnt}
+            edges[fnr2mnt]["value"] += 1
+
+
+        ctx = {
+            "nodes": list(nodes.values()),
+            "links": list(edges.values())
+        }
+
+
+        tmp = json.dumps(ctx, cls=NumpyEncoder)
+        tmp = json.loads(tmp)
+
+        self.report['sankey'] = tmp
+
+
+
+
+    def create_time_summary(self):
+        """
+        TODO: port to new object report
+
+        """
+
+        raise("Not implemented.")
+
+
+        # Original, Target:
+        ## <type>, <app time>, <read>, <write>, <meta>
+        #POSIX, 98.837925, 0.150075, 0.5991, 0.4129
+        #MPI-IO, 97.293875, 0.051575, 0.126525, 2.528025
+        #STDIO, 99.261425, 0, 0.738575, 0
+
+        # convienience links
+        summary = logdata['summary']
+        time_summary = logdata['time-summary.dat']
+
+        runtime = float(logdata['runtime'])
+        nprocs = int(logdata['nprocs'])
+
+        for layer in ['POSIX', 'MPIIO', 'STDIO']:
+            if (layer + '_OPENS') in summary or (layer + '_INDEP_OPENS') in summary :
+
+                entry = {
+                    'type': layer, 'app_time': None, 'read': None, 'write': None, 'meta': None
+                    }
+
+                io_time = 0.0
+                for op in ['READ', 'WRITE', 'META']:
+                    val = float(summary[layer + '_F_' + op + '_TIME'])
+                    io_time += val
+                    entry[op.lower()] = (val / (runtime * nprocs)) * 100
+
+                entry['app_time'] = ((runtime * nprocs - io_time) / (runtime * nprocs)) * 100
+                time_summary[layer] = entry
+
+
+
+
+
+
     def summarize(self):
         """
-        Compile summary 
+        Compiles a report summary of the records present in the report object.
+
+        Args:
+            None
+