Summary:
This is the Python 3.5+ version of our C API found in
include/mg_procedure.h. It should strive to be simple to use, but very
hard to misuse. The Python API avoids almost all memory management
issues found in the C API, so it is automatically less error prone.
Unfortunately, we cannot avoid the issue of having some objects being
alive only during the execution of a custom procedure. For such objects
we have a `Invalid<Object>Error` exception types. The Python API should
fairly easily check the validity because all such objects are related to
ProcCtx in some way. Our C/C++ part will push 1 instance of ProcCtx each
time a Python procedure is invoked and can therefore easily change
ProcCtx to be invalid. This can be as simple as changing the
implementation pointer of ProcCtx to NULL.
Reviewers: mferencevic, ipaljak, tlastre, dsantl, buda
Reviewed By: mferencevic, tlastre
Subscribers: pullbot
Differential Revision: https://phabricator.memgraph.io/D2650
Summary:
All external libraries now automatically include their include directories. It
is necessary only to link to the external library using
`target_link_libraries(target library)` and the include directory of the
library will be automatically available for the binary.
Reviewers: teon.banek
Reviewed By: teon.banek
Subscribers: pullbot
Differential Revision: https://phabricator.memgraph.io/D2654
Summary:
The antlr openCypher parser generation is moved to query. Also, header files
have been added to the list of generated files so that if any header file is
deleted CMake will know that it has to regenerate it.
Reviewers: teon.banek
Reviewed By: teon.banek
Subscribers: pullbot
Differential Revision: https://phabricator.memgraph.io/D2652
Summary:
The long running stress test had a subtle race condition which caused the test
to fail with an error message like "Runner X edges creation failed because of:
Can't serialize due to concurrent operations.". This situation was caused
because some workers could complete their initialization (initial vertex and
edge creation) before other workers. The workers that completed their
initialization would then proceed to execute the test. In the test they could
execute queries that make global updates on the graph that could interfere with
the concurrently running initialization queries of other workers.
This diff makes the runners wait until all initialization queries are fully
executed before they execute global operations on the graph.
Reviewers: teon.banek
Reviewed By: teon.banek
Subscribers: pullbot
Differential Revision: https://phabricator.memgraph.io/D2647
Summary:
This is an overview on how the example Query Module will look in
Python.
Reviewers: mferencevic, ipaljak, buda, tlastre, dsantl
Reviewed By: ipaljak, tlastre, dsantl
Subscribers: pullbot
Differential Revision: https://phabricator.memgraph.io/D2642
Summary:
The test is now more strict than before. Each verification step doesn't just
verify that object counts are correct, it now also verifies that all object IDs
are correct. The continuous integration script is improved to have a more
deterministic startup.
Reviewers: teon.banek, ipaljak
Reviewed By: teon.banek, ipaljak
Subscribers: pullbot
Differential Revision: https://phabricator.memgraph.io/D2606
Summary:
The property store stores a map of `PropertyId` to `PropertyValue` mappings. It
compresses all of the values in order to use as little memory as possible.
Reviewers: teon.banek, ipaljak
Reviewed By: teon.banek
Subscribers: buda, pullbot
Differential Revision: https://phabricator.memgraph.io/D2604
Summary:
This diff does not fix our Louvain implementation completely. The algorithm
internally still optimzies the wrong function, but this will at least correctly
calculate the modularity value at the end.
Fix of the algorithm will come in a separate diff
Reviewers: dsantl
Reviewed By: dsantl
Subscribers: pullbot
Differential Revision: https://phabricator.memgraph.io/D2603
