db.py revision 2343:a2b4a6ccee56 
1# Copyright (c) 2003-2004 The Regents of The University of Michigan
2# All rights reserved.
3#
4# Redistribution and use in source and binary forms, with or without
5# modification, are permitted provided that the following conditions are
6# met: redistributions of source code must retain the above copyright
7# notice, this list of conditions and the following disclaimer;
8# redistributions in binary form must reproduce the above copyright
9# notice, this list of conditions and the following disclaimer in the
10# documentation and/or other materials provided with the distribution;
11# neither the name of the copyright holders nor the names of its
12# contributors may be used to endorse or promote products derived from
13# this software without specific prior written permission.
14#
15# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
16# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
17# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
18# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
19# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
20# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
21# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
22# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
23# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
24# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
25# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26
27import MySQLdb, re, string
28
29def statcmp(a, b):
30    v1 = a.split('.')
31    v2 = b.split('.')
32
33    last = min(len(v1), len(v2)) - 1
34    for i,j in zip(v1[0:last], v2[0:last]):
35        if i != j:
36            return cmp(i, j)
37
38    # Special compare for last element.
39    if len(v1) == len(v2):
40        return cmp(v1[last], v2[last])
41    else:
42        return cmp(len(v1), len(v2))
43
44class RunData:
45    def __init__(self, row):
46        self.run = int(row[0])
47        self.name = row[1]
48        self.user = row[2]
49        self.project = row[3]
50
51class SubData:
52    def __init__(self, row):
53        self.stat = int(row[0])
54        self.x = int(row[1])
55        self.y = int(row[2])
56        self.name = row[3]
57        self.descr = row[4]
58
59class Data:
60    def __init__(self, row):
61        if len(row) != 5:
62            raise 'stat db error'
63        self.stat = int(row[0])
64        self.run = int(row[1])
65        self.x = int(row[2])
66        self.y = int(row[3])
67        self.data = float(row[4])
68
69    def __repr__(self):
70        return '''Data(['%d', '%d', '%d', '%d', '%f'])''' % ( self.stat,
71            self.run, self.x, self.y, self.data)
72
73class StatData(object):
74    def __init__(self, row):
75        self.stat = int(row[0])
76        self.name = row[1]
77        self.desc = row[2]
78        self.type = row[3]
79        self.prereq = int(row[5])
80        self.precision = int(row[6])
81
82        import flags
83        self.flags = 0
84        if int(row[4]): self.flags |= flags.printable
85        if int(row[7]): self.flags |= flags.nozero
86        if int(row[8]): self.flags |= flags.nonan
87        if int(row[9]): self.flags |= flags.total
88        if int(row[10]): self.flags |= flags.pdf
89        if int(row[11]): self.flags |= flags.cdf
90
91        if self.type == 'DIST' or self.type == 'VECTORDIST':
92            self.min = float(row[12])
93            self.max = float(row[13])
94            self.bktsize = float(row[14])
95            self.size = int(row[15])
96
97        if self.type == 'FORMULA':
98            self.formula = self.db.allFormulas[self.stat]
99
100class Node(object):
101    def __init__(self, name):
102        self.name = name
103    def __str__(self):
104        return self.name
105
106class Result(object):
107    def __init__(self, x, y):
108        self.data = {}
109        self.x = x
110        self.y = y
111
112    def __contains__(self, run):
113        return run in self.data
114
115    def __getitem__(self, run):
116        if run not in self.data:
117            self.data[run] = [ [ 0.0 ] * self.y for i in xrange(self.x) ]
118        return self.data[run]
119
120class Database(object):
121    def __init__(self):
122        self.host = 'zizzer.pool'
123        self.user = ''
124        self.passwd = ''
125        self.db = 'm5stats'
126        self.cursor = None
127
128        self.allStats = []
129        self.allStatIds = {}
130        self.allStatNames = {}
131
132        self.allSubData = {}
133
134        self.allRuns = []
135        self.allRunIds = {}
136        self.allRunNames = {}
137
138        self.allFormulas = {}
139
140        self.stattop = {}
141        self.statdict = {}
142        self.statlist = []
143
144        self.mode = 'sum';
145        self.runs = None
146        self.ticks = None
147        self.method = 'sum'
148        self._method = type(self).sum
149
150    def get(self, job, stat, system=None):
151        run = self.allRunNames.get(str(job), None)
152        if run is None:
153            return None
154
155        from info import ProxyError, scalar, vector, value, values, total, len
156        if system is None and hasattr(job, 'system'):
157            system = job.system
158
159        if system is not None:
160            stat.system = self[system]
161        try:
162            if scalar(stat):
163                return value(stat, run.run)
164            if vector(stat):
165                return values(stat, run.run)
166        except ProxyError:
167            return None
168
169        return None
170
171    def query(self, sql):
172        self.cursor.execute(sql)
173
174    def update_dict(self, dict):
175        dict.update(self.stattop)
176
177    def append(self, stat):
178        statname = re.sub(':', '__', stat.name)
179        path = string.split(statname, '.')
180        pathtop = path[0]
181        fullname = ''
182
183        x = self
184        while len(path) > 1:
185            name = path.pop(0)
186            if not x.__dict__.has_key(name):
187                x.__dict__[name] = Node(fullname + name)
188            x = x.__dict__[name]
189            fullname = '%s%s.' % (fullname, name)
190
191        name = path.pop(0)
192        x.__dict__[name] = stat
193
194        self.stattop[pathtop] = self.__dict__[pathtop]
195        self.statdict[statname] = stat
196        self.statlist.append(statname)
197
198    def connect(self):
199        # connect
200        self.thedb = MySQLdb.connect(db=self.db,
201                                     host=self.host,
202                                     user=self.user,
203                                     passwd=self.passwd)
204
205        # create a cursor
206        self.cursor = self.thedb.cursor()
207
208        self.query('''select rn_id,rn_name,rn_sample,rn_user,rn_project
209                   from runs''')
210        for result in self.cursor.fetchall():
211            run = RunData(result);
212            self.allRuns.append(run)
213            self.allRunIds[run.run] = run
214            self.allRunNames[run.name] = run
215
216        self.query('select sd_stat,sd_x,sd_y,sd_name,sd_descr from subdata')
217        for result in self.cursor.fetchall():
218            subdata = SubData(result)
219            if self.allSubData.has_key(subdata.stat):
220                self.allSubData[subdata.stat].append(subdata)
221            else:
222                self.allSubData[subdata.stat] = [ subdata ]
223
224        self.query('select * from formulas')
225        for id,formula in self.cursor.fetchall():
226            self.allFormulas[int(id)] = formula.tostring()
227
228        StatData.db = self
229        self.query('select * from stats')
230        import info
231        for result in self.cursor.fetchall():
232            stat = info.NewStat(self, StatData(result))
233            self.append(stat)
234            self.allStats.append(stat)
235            self.allStatIds[stat.stat] = stat
236            self.allStatNames[stat.name] = stat
237
238    # Name: listruns
239    # Desc: Prints all runs matching a given user, if no argument
240    #       is given all runs are returned
241    def listRuns(self, user=None):
242        print '%-40s %-10s %-5s' % ('run name', 'user', 'id')
243        print '-' * 62
244        for run in self.allRuns:
245            if user == None or user == run.user:
246                print '%-40s %-10s %-10d' % (run.name, run.user, run.run)
247
248    # Name: listTicks
249    # Desc: Prints all samples for a given run
250    def listTicks(self, runs=None):
251        print "tick"
252        print "----------------------------------------"
253        sql = 'select distinct dt_tick from data where dt_stat=1180 and ('
254        if runs != None:
255            first = True
256            for run in runs:
257               if first:
258            #       sql += ' where'
259                   first = False
260               else:
261                   sql += ' or'
262               sql += ' dt_run=%s' % run.run
263            sql += ')'
264        self.query(sql)
265        for r in self.cursor.fetchall():
266            print r[0]
267
268    # Name: retTicks
269    # Desc: Prints all samples for a given run
270    def retTicks(self, runs=None):
271        sql = 'select distinct dt_tick from data where dt_stat=1180 and ('
272        if runs != None:
273            first = True
274            for run in runs:
275               if first:
276                   first = False
277               else:
278                   sql += ' or'
279               sql += ' dt_run=%s' % run.run
280            sql += ')'
281        self.query(sql)
282        ret = []
283        for r in self.cursor.fetchall():
284            ret.append(r[0])
285        return ret
286
287    # Name: liststats
288    # Desc: Prints all statistics that appear in the database,
289    #         the optional argument is a regular expression that can
290    #         be used to prune the result set
291    def listStats(self, regex=None):
292        print '%-60s %-8s %-10s' % ('stat name', 'id', 'type')
293        print '-' * 80
294
295        rx = None
296        if regex != None:
297            rx = re.compile(regex)
298
299        stats = [ stat.name for stat in self.allStats ]
300        stats.sort(statcmp)
301        for stat in stats:
302            stat = self.allStatNames[stat]
303            if rx == None or rx.match(stat.name):
304                print '%-60s %-8s %-10s' % (stat.name, stat.stat, stat.type)
305
306    # Name: liststats
307    # Desc: Prints all statistics that appear in the database,
308    #         the optional argument is a regular expression that can
309    #         be used to prune the result set
310    def listFormulas(self, regex=None):
311        print '%-60s %s' % ('formula name', 'formula')
312        print '-' * 80
313
314        rx = None
315        if regex != None:
316            rx = re.compile(regex)
317
318        stats = [ stat.name for stat in self.allStats ]
319        stats.sort(statcmp)
320        for stat in stats:
321            stat = self.allStatNames[stat]
322            if stat.type == 'FORMULA' and (rx == None or rx.match(stat.name)):
323                print '%-60s %s' % (stat.name, self.allFormulas[stat.stat])
324
325    def getStat(self, stats):
326        if type(stats) is not list:
327            stats = [ stats ]
328
329        ret = []
330        for stat in stats:
331            if type(stat) is int:
332                ret.append(self.allStatIds[stat])
333
334            if type(stat) is str:
335                rx = re.compile(stat)
336                for stat in self.allStats:
337                    if rx.match(stat.name):
338                        ret.append(stat)
339        return ret
340
341    #########################################
342    # get the data
343    #
344    def query(self, op, stat, ticks, group=False):
345        sql = 'select '
346        sql += 'dt_stat as stat, '
347        sql += 'dt_run as run, '
348        sql += 'dt_x as x, '
349        sql += 'dt_y as y, '
350        if group:
351            sql += 'dt_tick as tick, '
352        sql += '%s(dt_data) as data ' % op
353        sql += 'from data '
354        sql += 'where '
355
356        if isinstance(stat, list):
357            val = ' or '.join([ 'dt_stat=%d' % s.stat for s in stat ])
358            sql += ' (%s)' % val
359        else:
360            sql += ' dt_stat=%d' % stat.stat
361
362        if self.runs != None and len(self.runs):
363            val = ' or '.join([ 'dt_run=%d' % r for r in self.runs ])
364            sql += ' and (%s)' % val
365
366        if ticks != None and len(ticks):
367            val = ' or '.join([ 'dt_tick=%d' % s for s in ticks ])
368            sql += ' and (%s)' % val
369
370        sql += ' group by dt_stat,dt_run,dt_x,dt_y'
371        if group:
372            sql += ',dt_tick'
373        return sql
374
375    # Name: sum
376    # Desc: given a run, a stat and an array of samples, total the samples
377    def sum(self, *args, **kwargs):
378        return self.query('sum', *args, **kwargs)
379
380    # Name: avg
381    # Desc: given a run, a stat and an array of samples, average the samples
382    def avg(self, stat, ticks):
383        return self.query('avg', *args, **kwargs)
384
385    # Name: stdev
386    # Desc: given a run, a stat and an array of samples, get the standard
387    #       deviation
388    def stdev(self, stat, ticks):
389        return self.query('stddev', *args, **kwargs)
390
391    def __setattr__(self, attr, value):
392        super(Database, self).__setattr__(attr, value)
393        if attr != 'method':
394            return
395
396        if value == 'sum':
397            self._method = self.sum
398        elif value == 'avg':
399            self._method = self.avg
400        elif value == 'stdev':
401            self._method = self.stdev
402        else:
403            raise AttributeError, "can only set get to: sum | avg | stdev"
404
405    def data(self, stat, ticks=None):
406        if ticks is None:
407            ticks = self.ticks
408        sql = self._method(self, stat, ticks)
409        self.query(sql)
410
411        runs = {}
412        xmax = 0
413        ymax = 0
414        for x in self.cursor.fetchall():
415            data = Data(x)
416            if not runs.has_key(data.run):
417                runs[data.run] = {}
418            if not runs[data.run].has_key(data.x):
419                runs[data.run][data.x] = {}
420
421            xmax = max(xmax, data.x)
422            ymax = max(ymax, data.y)
423            runs[data.run][data.x][data.y] = data.data
424
425        results = Result(xmax + 1, ymax + 1)
426        for run,data in runs.iteritems():
427            result = results[run]
428            for x,ydata in data.iteritems():
429                for y,data in ydata.iteritems():
430                    result[x][y] = data
431        return results
432
433    def __getitem__(self, key):
434        return self.stattop[key]
435