All python imports currently are added to the runfiles under __main__
which is only the default when no workspace name is provided. This
change supports both empty workspace names, and specified workspace
names.
This also supports python libs not anchored at root
`proto_library` targets are used as deps for `py_proto_library` and `py_grpc_library` rules. The `proto_library` targets can be configured using `import_prefix` and/or `strip_import_prefix` (which essentially move original location of the proto file and, as result, affects import path within proto files themselves and in the generated language-specific stubs).
The biggest question to answer when generating stubs from moved protos is where to put result (this decision affects all downstream rules as well, because the location of file affects its import path).
This PR tries to follow same logic as the native `cc_proto_library` (created and maintained by Bazel team). For example, if we have `firestore.proto` file, which is located under `google/firestore/v1beta1` Bazel package and want to move it under `google/cloud/firestore_v1beta1/proto` (this is axactly what happens in googleapis repository), it can be configured the following way:
```bzl
proto_library(
name = "firestore_moved_proto",
srcs = ["firestore.proto"],
import_prefix = "google/cloud/firestore_v1beta1/proto",
strip_import_prefix = "google/firestore/v1beta1",
)
```
The rule above will first generate virtual `.proto` files (under new location) and only after that generate a binary descriptor from them.
Specifically it will generate the following "virtual" file under `_virtual_imports` subdirectory of same package:
```
bazel-bin/google/firestore/v1beta1/_virtual_imports/firestore_moved_proto/google/cloud/firestore_v1beta1/proto/common.proto
```
When supplied to `cc_proto_library`, like the following:
```bzl
cc_proto_library(
name = "firestore_moved_cc_proto",
deps = ["firestore_moved_proto"],
)
```
The rule will generate .cc and .h files like the following:
```
bazel-bin/google/firestore/v1beta1/_virtual_imports/firestore_moved_proto/google/cloud/firestore_v1beta1/proto/firestore.pb.h
bazel-bin/google/firestore/v1beta1/_virtual_imports/firestore_moved_proto/google/cloud/firestore_v1beta1/proto/firestore.pb.cc
```
Notice, that each generated `.cc` and `.h` file is prefixed with `_virtual_imports/<name_of_proto_library_target>`.
The python rules try to do same thing, so for the following py_proto_library rule:
```bzl
py_proto_library(
name = "firestore_moved_py_proto",
deps = [":firestore_moved_proto"],
)
```
It wil generate the following file:
```
bazel-bin/google/firestore/v1beta1/_virtual_imports/firestore_moved_proto/google/cloud/firestore_v1beta1/proto/firestore_pb2.py
```
I.e in same path as cc_proto_library.
This all woks, but an annoying part is that to use the generated library in some other py_library, in needs to specify the "_virtual_imports/proto_target" path as its "includes" parameter (and I can't make it better than that).
Another option would be to skeep the `_virtual_imports/<name_of_proto_library_target>` the suffix for the generated python stubs, which will make the path look like the following:
```
bazel-bin/google/firestore/v1beta1/google/cloud/firestore_v1beta1/proto/firestore_pb2.py
```
That will make using generated stubs simpler and cleaner (no need to specify imports argument), but it will make it inconsistent with the other rules (like cc_proto_library) and also more susseptible to naming conflicts (if there is already something under the generated path).
By popular demand, we'll now be offering separate py_grpc_library and
py_proto_library targets sharing the same interface as within google3.
This change necessitated some modifications to how we pull in our own
Python-level dependencies and how we make those available to those
pulling in our project via Bazel.
There is now a grpc_python_deps() Bazel workspace rule that pulls in the
appropriate dependencies, which should be called from the client
project's WORKSPACE file. A test has been added to the bazel/test/
directory to verify that this behavior works as intended.
It's worth noting that the protobuf repository's usage of Starlark
bind() caused a great deal of trouble in ensuring that we could also
pull in six.
This change also required a change in the way generated proto code is
imported in the channelz and health-check modules, as well as in their
associated tests. We were importing them two different ways, each
relative. This resulted in two different module objects being imported
into the process, which were incompatible. I am not sure exactly what
caused this behavior to begin, as this should have been possible before
this PR. As a workaround, I am simply trying two different absolute
imports and using the one that works. This should function both inside
and outside of Bazel environments.
This commit resolves#18331.
This commit resolves#18256.
This commit resolves... another TODO that apparently didn't have an
associated github issue.
We swap out pubref's implementation of py_proto_library with our own,
which more closely mirrors the interface of the internal
py_proto_library, taking the descriptor file output of a proto_library
rule as input.
One minor change in behavior was introduced for simplicity. When a
py_proto_library depends on a proto_library with a source proto file in
a subdirectory of the bazel package, the import module of the resultant
python library will reflect the package, *not* the full directory of the
proto file, including both the bazel package and the subdirectories, as
pubref did previously. This behavior also more closely mirrors google
internal behavior.
This commit also introduces a slightly more stringent bazel format
script. Buildifier on its own will not take care of long lines, but by
running yapf first, we end up with a more legible file. At the moment,
there is no sanity check associated with this formatter.
Matt Wilson
Yannic Bonenberger
Lidi Zheng
vam-google
Richard Belleville
Juanli Shen
Esun Kim
Marcel Hlopko
Keith Smiley