10
This document describes the indexing facilities within bzrlib.
19
To provide a clean concept of index that can be reused by different
20
components within the codebase rather than being rewritten every time
21
by different components.
27
An **index** is a dictionary mapping opaque keys to opaque values.
28
Different index types may allow some of the value data to be interpreted
29
by the index. For example the ``GraphIndex`` index stores a graph between
30
keys as part of the index.
36
bzr is moving to a write-once model for repository storage in order to
37
achieve lock-free repositories eventually. In order to support this, we are
38
making our new index classes **immutable**. That is, one creates a new
39
index in a single operation, and after that it is read only. To combine
40
two indices a ``Combined*`` index may be used, or an **index merge** may
41
be performed by reading the entire value of two (or more) indices and
42
writing them into a new index.
47
We may end up with multiple different Index types (e.g. GraphIndex,
48
Index, WhackyIndex). Even though these may require different method
49
signatures to operate would strive to keep the signatures and return
50
values as similar as possible. e.g.::
52
GraphIndexBuilder - add_node(key, value, references)
53
IndexBuilder - add_node(key, value)
54
WhackyIndexBuilder - add_node(key, value, whackiness)
56
as opposed to something quite different like::
58
node = IncrementalBuilder.get_node()
65
An initial implementation of indexing can probably get away with a small
66
number of primitives. Assuming we have write once index files:
71
This should be done by creating an ``IndexBuilder`` and then calling
72
``insert(key, value)`` many times. (Indices that support sorting,
73
topological sorting etc, will want specialised insert methods).
75
When the keys have all been added, a ``finish`` method should be called,
76
which will return a file stream to read the index data from.
78
Retrieve entries from the index
79
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
81
This should allow random access to the index using readv, so we probably
82
want to open the index on a ``Transport``, then use ``iter_entries(keys)``,
83
which can return an iterator that yields ``(key, value)`` pairs in
84
whatever order makes sense for the index.
89
Merging of N indices requires a concordance of the keys of the index. So
90
we should offer a ``iter_all_entries`` call that has the same return type
91
as the ``iter_entries`` call.
99
``GraphIndex`` supports graph based lookups. While currently unoptimised
100
for reading, the index is quite space efficient at storing the revision
101
graph index for bzr. The ``GraphIndexBuilder`` may be used to create one
102
of these indices by calling ``add_node`` until all nodes are added, then
103
``finish`` to obtain a file stream containing the index data. Multiple
104
indices may be queried using the ``CombinedGraphIndex`` class.