comparison query.cpp @ 469:d3afc91d205d api-inversion

Move audioDB::query over to audioDB.cpp

At the same time, remove all the abstraction violations in
audioDB::query, which came in two flavours: use of dbH->numFiles, which
is dealt with by getting the database status instead (and is eventually
unnecessary, being only needed now because reporters are implemented in
terms of vectors indexed by ID), and use of fileTable in reporter's
report functions (dealt with by passing in the adb instead).

To actually implement reporting as of now, we continue to use stuff from
audioDB-internals.h; maybe someday we will be clean and shiny.

 #include "audioDB.h"
-#include "reporter.h"
-
 #include "audioDB-internals.h"
 #include "accumulators.h"
 
 bool audiodb_powers_acceptable(adb_query_refine_t *r, double p1, double p2) {
   if (r->flags & ADB_REFINE_ABSOLUTE_THRESHOLD) {

     }
   }
   return true;
 }
 
-void audioDB::query(const char* dbName, const char* inFile, adb__queryResponse *adbQueryResponse) {
-
-  // init database tables and dbH first
-  if(query_from_key)
-    initTables(dbName);
-  else
-    initTables(dbName, inFile);
-
-  adb_query_spec_t qspec;
-  adb_datum_t datum = {0};
-
-  qspec.refine.flags = 0;
-  if(trackFile) {
-    qspec.refine.flags |= ADB_REFINE_INCLUDE_KEYLIST;
-    std::vector<const char *> v;
-    char *k = new char[MAXSTR];
-    trackFile->getline(k, MAXSTR);    
-    while(!trackFile->eof()) {
-      v.push_back(k);
-      k = new char[MAXSTR];
-      trackFile->getline(k, MAXSTR);    
-    }
-    delete [] k;
-    qspec.refine.include.nkeys = v.size();
-    qspec.refine.include.keys = new const char *[qspec.refine.include.nkeys];
-    for(unsigned int k = 0; k < qspec.refine.include.nkeys; k++) {
-      qspec.refine.include.keys[k] = v[k];
-    }
-  }
-  if(query_from_key) {
-    qspec.refine.flags |= ADB_REFINE_EXCLUDE_KEYLIST;
-    qspec.refine.exclude.nkeys = 1;
-    qspec.refine.exclude.keys = &key;
-  }
-  if(radius) {
-    qspec.refine.flags |= ADB_REFINE_RADIUS;
-    qspec.refine.radius = radius;
-  }
-  if(use_absolute_threshold) {
-    qspec.refine.flags |= ADB_REFINE_ABSOLUTE_THRESHOLD;
-    qspec.refine.absolute_threshold = absolute_threshold;
-  }
-  if(use_relative_threshold) {
-    qspec.refine.flags |= ADB_REFINE_RELATIVE_THRESHOLD;
-    qspec.refine.relative_threshold = relative_threshold;
-  }
-  if(usingTimes) {
-    qspec.refine.flags |= ADB_REFINE_DURATION_RATIO;
-    qspec.refine.duration_ratio = timesTol;
-  }
-  /* FIXME: not sure about this any more; maybe it belongs in
-     query_id?  Or maybe we just don't need a flag for it? */
-  qspec.refine.hopsize = sequenceHop;
-  if(sequenceHop != 1) {
-    qspec.refine.flags |= ADB_REFINE_HOP_SIZE;
-  }
-
-  if(query_from_key) {
-    datum.key = key;
-  } else {
-    int fd;
-    struct stat st;
-
-    /* FIXME: around here there are all sorts of hideous leaks. */
-    fd = open(inFile, O_RDONLY);
-    if(fd < 0) {
-      error("failed to open feature file", inFile);
-    }
-    fstat(fd, &st);
-    read(fd, &datum.dim, sizeof(uint32_t));
-    datum.nvectors = (st.st_size - sizeof(uint32_t)) / (datum.dim * sizeof(double));
-    datum.data = (double *) malloc(st.st_size - sizeof(uint32_t));
-    read(fd, datum.data, st.st_size - sizeof(uint32_t));
-    close(fd);
-    if(usingPower) {
-      uint32_t one;
-      fd = open(powerFileName, O_RDONLY);
-      if(fd < 0) {
-        error("failed to open power file", powerFileName);
-      }
-      read(fd, &one, sizeof(uint32_t));
-      if(one != 1) {
-        error("malformed power file dimensionality", powerFileName);
-      }
-      datum.power = (double *) malloc(datum.nvectors * sizeof(double));
-      if(read(fd, datum.power, datum.nvectors * sizeof(double)) != (ssize_t) (datum.nvectors * sizeof(double))) {
-        error("malformed power file", powerFileName);
-      }
-      close(fd);
-    }
-    if(usingTimes) {
-      datum.times = (double *) malloc(2 * datum.nvectors * sizeof(double));
-      insertTimeStamps(datum.nvectors, timesFile, datum.times);
-    }
-  }
-
-  qspec.qid.datum = &datum;
-  qspec.qid.sequence_length = sequenceLength;
-  qspec.qid.flags = 0;
-  qspec.qid.flags |= usingQueryPoint ? 0 : ADB_QID_FLAG_EXHAUSTIVE;
-  qspec.qid.flags |= lsh_exact ? 0 : ADB_QID_FLAG_ALLOW_FALSE_POSITIVES;
-  qspec.qid.sequence_start = queryPoint;
-
-  switch(queryType) {
-  case O2_POINT_QUERY:
-    qspec.qid.sequence_length = 1;
-    qspec.params.accumulation = ADB_ACCUMULATION_DB;
-    qspec.params.distance = ADB_DISTANCE_DOT_PRODUCT;
-    qspec.params.npoints = pointNN;
-    qspec.params.ntracks = 0;
-    reporter = new pointQueryReporter< std::greater < NNresult > >(pointNN);
-    break;
-  case O2_TRACK_QUERY:
-    qspec.qid.sequence_length = 1;
-    qspec.params.accumulation = ADB_ACCUMULATION_PER_TRACK;
-    qspec.params.distance = ADB_DISTANCE_DOT_PRODUCT;
-    qspec.params.npoints = pointNN;
-    qspec.params.ntracks = trackNN;
-    reporter = new trackAveragingReporter< std::greater< NNresult > >(pointNN, trackNN, dbH->numFiles);
-    break;
-  case O2_SEQUENCE_QUERY:
-  case O2_N_SEQUENCE_QUERY:
-    qspec.params.accumulation = ADB_ACCUMULATION_PER_TRACK;
-    qspec.params.distance = no_unit_norming ? ADB_DISTANCE_EUCLIDEAN : ADB_DISTANCE_EUCLIDEAN_NORMED;
-    qspec.params.npoints = pointNN;
-    qspec.params.ntracks = trackNN;
-    switch(queryType) {
-    case O2_SEQUENCE_QUERY:
-      if(!(qspec.refine.flags & ADB_REFINE_RADIUS)) {
-        reporter = new trackAveragingReporter< std::less< NNresult > >(pointNN, trackNN, dbH->numFiles);
-      } else {
-	reporter = new trackSequenceQueryRadReporter(trackNN, dbH->numFiles);
-      }
-      break;
-    case O2_N_SEQUENCE_QUERY:
-      if(!(qspec.refine.flags & ADB_REFINE_RADIUS)) {
-        reporter = new trackSequenceQueryNNReporter< std::less < NNresult > >(pointNN, trackNN, dbH->numFiles);
-      } else {
-	reporter = new trackSequenceQueryRadNNReporter(pointNN, trackNN, dbH->numFiles);
-      }
-      break;
-    }
-    break;
-  case O2_ONE_TO_ONE_N_SEQUENCE_QUERY:
-    qspec.params.accumulation = ADB_ACCUMULATION_ONE_TO_ONE;
-    qspec.params.distance = ADB_DISTANCE_EUCLIDEAN_NORMED;
-    qspec.params.npoints = 0;
-    qspec.params.ntracks = 0;
-    break;
-  default:
-    error("unrecognized queryType");
-  }
-
-  /* Somewhere around here is where the implementation of
-   * audiodb_query_spec() starts. */
-
-  adb_qstate_internal_t qstate;
+adb_query_results_t *audiodb_query_spec(adb_t *adb, adb_query_spec_t *qspec) {
+  adb_qstate_internal_t qstate = {0};
   qstate.allowed_keys = new std::set<std::string>;
-  if(qspec.refine.flags & ADB_REFINE_INCLUDE_KEYLIST) {
-    for(unsigned int k = 0; k < qspec.refine.include.nkeys; k++) {
-      qstate.allowed_keys->insert(qspec.refine.include.keys[k]);
+  adb_query_results_t *results;
+  if(qspec->refine.flags & ADB_REFINE_INCLUDE_KEYLIST) {
+    for(unsigned int k = 0; k < qspec->refine.include.nkeys; k++) {
+      qstate.allowed_keys->insert(qspec->refine.include.keys[k]);
     }
   } else {
     for(unsigned int k = 0; k < adb->header->numFiles; k++) {
       qstate.allowed_keys->insert((*adb->keys)[k]);
     }
   }
-  if(qspec.refine.flags & ADB_REFINE_EXCLUDE_KEYLIST) {
-    for(unsigned int k = 0; k < qspec.refine.exclude.nkeys; k++) {
-      qstate.allowed_keys->erase(qspec.refine.exclude.keys[k]);
-    }
-  }
-
-  switch(qspec.params.distance) {
+  if(qspec->refine.flags & ADB_REFINE_EXCLUDE_KEYLIST) {
+    for(unsigned int k = 0; k < qspec->refine.exclude.nkeys; k++) {
+      qstate.allowed_keys->erase(qspec->refine.exclude.keys[k]);
+    }
+  }
+
+  switch(qspec->params.distance) {
   case ADB_DISTANCE_DOT_PRODUCT:
-    switch(qspec.params.accumulation) {
+    switch(qspec->params.accumulation) {
     case ADB_ACCUMULATION_DB:
-      qstate.accumulator = new DBAccumulator<adb_result_dist_gt>(qspec.params.npoints);
+      qstate.accumulator = new DBAccumulator<adb_result_dist_gt>(qspec->params.npoints);
       break;
     case ADB_ACCUMULATION_PER_TRACK:
-      qstate.accumulator = new PerTrackAccumulator<adb_result_dist_gt>(qspec.params.npoints, qspec.params.ntracks);
+      qstate.accumulator = new PerTrackAccumulator<adb_result_dist_gt>(qspec->params.npoints, qspec->params.ntracks);
       break;
     case ADB_ACCUMULATION_ONE_TO_ONE:
       qstate.accumulator = new NearestAccumulator<adb_result_dist_gt>();
       break;
     default:
-      error("unknown accumulation");
+      goto error;
     }
     break;
   case ADB_DISTANCE_EUCLIDEAN_NORMED:
   case ADB_DISTANCE_EUCLIDEAN:
-    switch(qspec.params.accumulation) {
+    switch(qspec->params.accumulation) {
     case ADB_ACCUMULATION_DB:
-      qstate.accumulator = new DBAccumulator<adb_result_dist_lt>(qspec.params.npoints);
+      qstate.accumulator = new DBAccumulator<adb_result_dist_lt>(qspec->params.npoints);
       break;
     case ADB_ACCUMULATION_PER_TRACK:
-      qstate.accumulator = new PerTrackAccumulator<adb_result_dist_lt>(qspec.params.npoints, qspec.params.ntracks);
+      qstate.accumulator = new PerTrackAccumulator<adb_result_dist_lt>(qspec->params.npoints, qspec->params.ntracks);
       break;
     case ADB_ACCUMULATION_ONE_TO_ONE:
       qstate.accumulator = new NearestAccumulator<adb_result_dist_lt>();
       break;
     default:
-      error("unknown accumulation");
+      goto error;
     }
     break;
   default:
-    error("unknown distance function");
+    goto error;
   }
   
-  // Test for index (again) here
-  if((qspec.refine.flags & ADB_REFINE_RADIUS) && audiodb_index_exists(adb->path, qspec.refine.radius, qspec.qid.sequence_length)){ 
-    VERB_LOG(1, "Calling indexed query on database %s, radius=%f, sequence_length=%d\n", adb->path, qspec.refine.radius, qspec.qid.sequence_length);
-    if(audiodb_index_query_loop(adb, &qspec, &qstate) < 0) {
-      error("index_query_loop failed");
+  if((qspec->refine.flags & ADB_REFINE_RADIUS) && audiodb_index_exists(adb->path, qspec->refine.radius, qspec->qid.sequence_length)) {
+    if(audiodb_index_query_loop(adb, qspec, &qstate) < 0) {
+      goto error;
     }
   } else {
-    VERB_LOG(1, "Calling brute-force query on database %s\n", dbName);
-    if(audiodb_query_loop(adb, &qspec, &qstate)) {
-      error("audiodb_query_loop failed");
-    }
-  }
-
-  adb_query_results_t *rs = qstate.accumulator->get_points();
+    if(audiodb_query_loop(adb, qspec, &qstate)) {
+      goto error;
+    }
+  }
+
+  results = qstate.accumulator->get_points();
 
   delete qstate.accumulator;
   delete qstate.allowed_keys;
 
-  /* End of audiodb_query_spec() function */
-
-  for(unsigned int k = 0; k < rs->nresults; k++) {
-    adb_result_t r = rs->results[k];
-    reporter->add_point(audiodb_key_index(adb, r.key), r.qpos, r.ipos, r.dist);
-  }
-  audiodb_query_free_results(adb, &qspec, rs);
-
-  reporter->report(fileTable, adbQueryResponse);
+  return results;
+
+ error:
+  if(qstate.accumulator)
+    delete qstate.accumulator;
+  if(qstate.allowed_keys)
+    delete qstate.allowed_keys;
+  return NULL;
 }
 
 int audiodb_query_free_results(adb_t *adb, adb_query_spec_t *spec, adb_query_results_t *rs) {
   free(rs->results);
   free(rs);

   read_or_goto_error(trkfid, *data_buffer_p, track_size);
   return 0;
 
  error:
   return 1;
-}
-
-void audioDB::insertTimeStamps(unsigned numVectors, std::ifstream *timesFile, double *timesdata) {
-  assert(usingTimes);
-
-  unsigned numtimes = 0;
-
-  if(!timesFile->is_open()) {
-    error("problem opening times file on timestamped database", timesFileName);
-  }
-
-  double timepoint, next;
-  *timesFile >> timepoint;
-  if (timesFile->eof()) {
-    error("no entries in times file", timesFileName);
-  }
-  numtimes++;
-  do {
-    *timesFile >> next;
-    if (timesFile->eof()) {
-      break;
-    }
-    numtimes++;
-    timesdata[0] = timepoint;
-    timepoint = (timesdata[1] = next);
-    timesdata += 2;
-  } while (numtimes < numVectors + 1);
-
-  if (numtimes < numVectors + 1) {
-    error("too few timepoints in times file", timesFileName);
-  }
-
-  *timesFile >> next;
-  if (!timesFile->eof()) {
-    error("too many timepoints in times file", timesFileName);
-  }
 }
 
 int audiodb_track_id_datum(adb_t *adb, uint32_t track_id, adb_datum_t *d) {
   off_t track_offset = (*adb->track_offsets)[track_id];
   if(adb->header->flags & O2_FLAG_LARGE_ADB) {