Skip to content

Conversation

@yongtang
Copy link
Member

@yongtang yongtang commented Aug 1, 2019

This is part of the effort to reduce the dedicated C++
implementation of Dataset and replace with primitive ops
that could be used both with tf.data, and with Tensor.

See #382 and #366 for related discussions.

There are some room for enhancement, for example,
a timestamp could be passed to read_prometheus
and each call will only read a small slice of the data
to tensor. Will have follow up PRs later to implement that.

Signed-off-by: Yong Tang [email protected]

This is part of the effort to reduce the dedicated C++
implementation of Dataset and replace with primitive ops
that could be used both with tf.data, and with Tensor.

There are some room for enhancement, for example,
a timestamp could be passed to read_prometheus
and each call will only read a small slice of the data
to tensor. Will have follow up PRs later to implement that.

Signed-off-by: Yong Tang <[email protected]>
@yongtang
Copy link
Member Author

yongtang commented Aug 4, 2019

Plan to merge this PR shortly.

@yongtang yongtang merged commit c855890 into tensorflow:master Aug 4, 2019
@yongtang yongtang deleted the prometheus branch August 4, 2019 16:52
i-ony pushed a commit to i-ony/io that referenced this pull request Feb 8, 2021
This is part of the effort to reduce the dedicated C++
implementation of Dataset and replace with primitive ops
that could be used both with tf.data, and with Tensor.

There are some room for enhancement, for example,
a timestamp could be passed to read_prometheus
and each call will only read a small slice of the data
to tensor. Will have follow up PRs later to implement that.

Signed-off-by: Yong Tang <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant