WIP: Slicing and refactoring offline

Closes #37 (closed)

km3io/offline.py

 from collections import namedtuple
 import uproot
 import numpy as np
-import awkward1 as ak
 import warnings
 from .definitions import mc_header
 
 MAIN_TREE_NAME = "E"
 # 110 MB based on the size of the largest basket found so far in km3net
 BASKET_CACHE_SIZE = 110 * 1024**2
+BASKET_CACHE = uproot.cache.ThreadSafeArrayCache(BASKET_CACHE_SIZE)
+EXCLUDE_KEYS = set(["AAObject", "t", "fBits", "fUniqueID"])
 
 BranchMapper = namedtuple(
     "BranchMapper",
-    ['name', 'key', 'extra', 'exclude', 'update', 'attrparser', 'flat'])
+    ['name', 'key', 'extra', 'exclude', 'update', 'attrparser'])
 
 
 def _nested_mapper(key):
@@ -19,26 +20,57 @@ def _nested_mapper(key):
     return '_'.join(key.split('.')[1:])
 
 
-EXCLUDE_KEYS = set(["AAObject", "t", "fBits", "fUniqueID"])
-BRANCH_MAPS = [
-    BranchMapper("tracks", "trks", {}, ['trks.usr_data', 'trks.usr'], {},
-                 _nested_mapper, False),
-    BranchMapper("mc_tracks", "mc_trks", {},
-                 ['mc_trks.usr_data', 'mc_trks.usr'], {}, _nested_mapper,
-                 False),
-    BranchMapper("hits", "hits", {}, ['hits.usr'], {}, _nested_mapper, False),
-    BranchMapper("mc_hits", "mc_hits", {},
-                 ['mc_hits.usr', 'mc_hits.dom_id', 'mc_hits.channel_id'], {},
-                 _nested_mapper, False),
-    BranchMapper("events", "Evt", {
-        't_sec': 't.fSec',
-        't_ns': 't.fNanoSec'
-    }, [], {
-        'n_hits': 'hits',
-        'n_mc_hits': 'mc_hits',
-        'n_tracks': 'trks',
-        'n_mc_tracks': 'mc_trks'
-    }, lambda a: a, True),
+EVENTS_MAP = BranchMapper(name="events",
+                          key="Evt",
+                          extra={
+                              't_sec': 't.fSec',
+                              't_ns': 't.fNanoSec'
+                          },
+                          exclude=[],
+                          update={
+                              'n_hits': 'hits',
+                              'n_mc_hits': 'mc_hits',
+                              'n_tracks': 'trks',
+                              'n_mc_tracks': 'mc_trks'
+                          },
+                          attrparser=lambda a: a)
+
+SUBBRANCH_MAPS = [
+    BranchMapper(
+        name="tracks",
+        key="trks",
+        extra={},
+        exclude=['trks.usr_data', 'trks.usr', 'trks.fUniqueID', 'trks.fBits'],
+        update={},
+        attrparser=_nested_mapper),
+    BranchMapper(name="mc_tracks",
+                 key="mc_trks",
+                 extra={},
+                 exclude=[
+                     'mc_trks.usr_data', 'mc_trks.usr', 'mc_trks.rec_stages',
+                     'mc_trks.fitinf', 'mc_trks.fUniqueID', 'mc_trks.fBits'
+                 ],
+                 update={},
+                 attrparser=_nested_mapper),
+    BranchMapper(name="hits",
+                 key="hits",
+                 extra={},
+                 exclude=[
+                     'hits.usr', 'hits.pmt_id', 'hits.origin', 'hits.a',
+                     'hits.pure_a', 'hits.fUniqueID', 'hits.fBits'
+                 ],
+                 update={},
+                 attrparser=_nested_mapper),
+    BranchMapper(name="mc_hits",
+                 key="mc_hits",
+                 extra={},
+                 exclude=[
+                     'mc_hits.usr', 'mc_hits.dom_id', 'mc_hits.channel_id',
+                     'mc_hits.tdc', 'mc_hits.tot', 'mc_hits.trig',
+                     'mc_hits.fUniqueID', 'mc_hits.fBits'
+                 ],
+                 update={},
+                 attrparser=_nested_mapper),
 ]
 
 
@@ -56,11 +88,7 @@ class cached_property:
 
 class OfflineReader:
     """reader for offline ROOT files"""
-    def __init__(self,
-                 file_path=None,
-                 fobj=None,
-                 data=None,
-                 index=slice(None)):
+    def __init__(self, file_path=None, fobj=None, data=None, index=None):
         """ OfflineReader class is an offline ROOT file wrapper
 
         Parameters
@@ -74,17 +102,18 @@ class OfflineReader:
         if file_path is not None:
             self._fobj = uproot.open(file_path)
             self._tree = self._fobj[MAIN_TREE_NAME]
-            self._data = self._tree.lazyarrays(
-                basketcache=uproot.cache.ThreadSafeArrayCache(
-                    BASKET_CACHE_SIZE))
+            self._data = self._tree.lazyarrays(basketcache=BASKET_CACHE)
         else:
             self._fobj = fobj
             self._tree = self._fobj[MAIN_TREE_NAME]
             self._data = data
 
-        for mapper in BRANCH_MAPS:
-            setattr(self, mapper.name,
-                    Branch(self._tree, mapper=mapper, index=self._index))
+    @cached_property
+    def events(self):
+        return Branch(self._tree,
+                      mapper=EVENTS_MAP,
+                      index=self._index,
+                      subbranchmaps=SUBBRANCH_MAPS)
 
     @classmethod
     def from_index(cls, source, index):
@@ -107,12 +136,10 @@ class OfflineReader:
 
     def __len__(self):
         tree = self._fobj[MAIN_TREE_NAME]
-        if self._index == slice(None):
+        if self._index is None:
             return len(tree)
         else:
-            return len(
-                tree.lazyarrays(basketcache=uproot.cache.ThreadSafeArrayCache(
-                    BASKET_CACHE_SIZE))[self.index])
+            return len(tree.lazyarrays(basketcache=BASKET_CACHE)[self.index])
 
     @cached_property
     def header(self):
@@ -418,10 +445,10 @@ class OfflineReader:
             are not found, None is returned as the stages index.
         """
         if mc is False:
-            stages_data = self.tracks.rec_stages
+            stages_data = self.events.tracks.rec_stages
 
         if mc is True:
-            stages_data = self.mc_tracks.rec_stages
+            stages_data = self.events.mc_tracks.rec_stages
 
         for trk_index, rec_stages in enumerate(stages_data):
             try:
@@ -465,18 +492,19 @@ class OfflineReader:
 
 
 class Usr:
-    """Helper class to access AAObject usr stuff"""
-    def __init__(self, name, tree, index=slice(None)):
+    """Helper class to access AAObject `usr`` stuff"""
+    def __init__(self, name, tree, index=None):
         # Here, we assume that every event has the same names in the same order
         # to massively increase the performance. This needs triple check if it's
         # always the case; the usr-format is simply a very bad design.
         self._name = name
         try:
             tree['usr']  # This will raise a KeyError in old aanet files
-                         # which has a different strucuter and key (usr_data)
-                         # We do not support those...
+            # which has a different strucuter and key (usr_data)
+            # We do not support those...
             self._usr_names = [
-                n.decode("utf-8") for n in tree['usr_names'].array()[0]
+                n.decode("utf-8") for n in tree['usr_names'].lazyarray(
+                    basketcache=BASKET_CACHE)[0]
             ]
         except (KeyError, IndexError):  # e.g. old aanet files
             self._usr_names = []
@@ -485,9 +513,10 @@ class Usr:
                 name: index
                 for index, name in enumerate(self._usr_names)
             }
-            self._usr_data = tree['usr'].lazyarray(
-                basketcache=uproot.cache.ThreadSafeArrayCache(
-                    BASKET_CACHE_SIZE))[index]
+            data = tree['usr'].lazyarray(basketcache=BASKET_CACHE)
+            if index is not None:
+                data = data[index]
+            self._usr_data = data
             for name in self._usr_names:
                 setattr(self, name, self[name])
 
@@ -523,7 +552,7 @@ def _to_num(value):
 
 
 class Header:
-    """The online header"""
+    """The header"""
     def __init__(self, header):
         self._data = {}
         for attribute, fields in mc_header.items():
@@ -549,17 +578,36 @@ class Header:
 
 class Branch:
     """Branch accessor class"""
-    def __init__(self, tree, mapper, index=slice(None)):
+    def __init__(self,
+                 tree,
+                 mapper,
+                 index=None,
+                 subbranchmaps=None,
+                 keymap=None):
         self._tree = tree
         self._mapper = mapper
         self._index = index
         self._keymap = None
         self._branch = tree[mapper.key]
+        self._subbranches = []
 
-        self._initialise_keys()
+        if keymap is None:
+            self._initialise_keys()  #
+        else:
+            self._keymap = keymap
+
+        if subbranchmaps is not None:
+            for mapper in subbranchmaps:
+                subbranch = Branch(self._tree,
+                                   mapper=mapper,
+                                   index=self._index)
+                self._subbranches.append(subbranch)
+        for subbranch in self._subbranches:
+            setattr(self, subbranch._mapper.name, subbranch)
 
     def _initialise_keys(self):
-        """Create the keymap and instance attributes"""
+        """Create the keymap and instance attributes for branch keys"""
+        # TODO: this could be a cached property
         keys = set(k.decode('utf-8') for k in self._branch.keys()) - set(
             self._mapper.exclude) - EXCLUDE_KEYS
         self._keymap = {
@@ -571,11 +619,8 @@ class Branch:
         for k in self._mapper.update.values():
             del self._keymap[k]
 
-        # self._EntryType = namedtuple(mapper.name[:-1], self.keys())
-
-        for key in self.keys():
-            # print("setting", self._mapper.name, key)
-            setattr(self, key, self[key])
+        for key in self._keymap.keys():
+            setattr(self, key, None)
 
     def keys(self):
         return self._keymap.keys()
@@ -584,75 +629,51 @@ class Branch:
     def usr(self):
         return Usr(self._mapper.name, self._branch, index=self._index)
 
+    def __getattribute__(self, attr):
+        if attr.startswith("_"):  # let all private and magic methods pass
+            return object.__getattribute__(self, attr)
+
+        if attr in self._keymap.keys():  # intercept branch key lookups
+            out = self._branch[self._keymap[attr]].lazyarray(
+                basketcache=BASKET_CACHE)
+            if self._index is not None:
+                out = out[self._index]
+            return out
+
+        return object.__getattribute__(self, attr)
+
     def __getitem__(self, item):
-        """Slicing magic a la numpy"""
-        if isinstance(item, slice):
-            return self.__class__(self._tree, self._mapper, index=item)
-        if isinstance(item, int):
-            if self._mapper.flat:
-                return BranchElement(
-                    self._mapper.name, {
-                        key:
-                        self._branch[self._keymap[key]].array()[self._index]
-                        for key in self.keys()
-                    })[item]
-            else:
-                return BranchElement(
-                    self._mapper.name, {
-                        key:
-                        self._branch[self._keymap[key]].array()[self._index,
-                                                                item]
-                        for key in self.keys()
-                    })
+        """Slicing magic"""
+        if isinstance(item, (int, slice)):
+            return self.__class__(self._tree,
+                                  self._mapper,
+                                  index=item,
+                                  keymap=self._keymap,
+                                  subbranchmaps=SUBBRANCH_MAPS)
+
         if isinstance(item, tuple):
             return self[item[0]][item[1]]
-        return self._branch[self._keymap[item]].lazyarray(
-            basketcache=uproot.cache.ThreadSafeArrayCache(BASKET_CACHE_SIZE))[
-                self._index]
+
+        return self.__class__(self._tree,
+                              self._mapper,
+                              index=np.array(item),
+                              keymap=self._keymap,
+                              subbranchmaps=SUBBRANCH_MAPS)
 
     def __len__(self):
-        if self._index == slice(None):
+        if self._index is None:
             return len(self._branch)
+        elif isinstance(self._index, int):
+            return 1
         else:
-            return len(
-                self._branch[self._keymap['id']].lazyarray()[self._index])
+            return len(self._branch[self._keymap['id']].lazyarray(
+                basketcache=BASKET_CACHE)[self._index])
 
     def __str__(self):
         return "Number of elements: {}".format(len(self._branch))
 
     def __repr__(self):
-        return "<{}[{}]: {} parsed elements>".format(self.__class__.__name__,
-                                                     self._mapper.name,
-                                                     len(self))
-
-
-class BranchElement:
-    """Represents a single branch element
-
-    Parameters
-    ----------
-    name: str
-        The name of the branch
-    dct: dict (keys=attributes, values=arrays of values)
-        The data
-    index: slice
-        The slice mask to be applied to the sub-arrays
-    """
-    def __init__(self, name, dct, index=slice(None)):
-        self._dct = dct
-        self._name = name
-        self._index = index
-        self.ItemConstructor = namedtuple(self._name[:-1], dct.keys())
-        for key, values in dct.items():
-            setattr(self, key, values[index])
-
-    def __getitem__(self, item):
-        if isinstance(item, slice):
-            return self.__class__(self._name, self._dct, index=item)
-        if isinstance(item, int):
-            return self.ItemConstructor(
-                **{k: v[self._index][item]
-                   for k, v in self._dct.items()})
-
-    def __repr__(self):
-        return "<{}[{}]>".format(self.__class__.__name__, self._name)
+        length = len(self)
+        return "<{}[{}]: {} element{}>".format(self.__class__.__name__,
+                                               self._mapper.name, length,
+                                               's' if length > 1 else '')