# Suite of verification tests for the SINGLE TensorFlow wheel in /tf/pkg
# or whatever path is set as $TF_WHEEL.

setup_file() {
    cd /tf/pkg
    if [[ -z "$TF_WHEEL" ]]; then
        export TF_WHEEL=$(find /tf/pkg -iname "*.whl")
    fi
}

teardown_file() {
    rm -rf /tf/venv
}

@test "Wheel is manylinux2014 (manylinux_2_17) compliant" {
    python3 -m auditwheel show "$TF_WHEEL" > audit.txt

  }

  @Test fun serverHelloTwoChunks() {
    data.write("818537fa213d7f9f4d".decodeHex()) // Hel
    data.write("5158".decodeHex()) // lo
    serverReader.processNextFrame()
    callback.assertTextMessage("Hello")
  }

  @Test fun serverWithCompressionHelloTwoChunks() {
    data.write("818537fa213d7f9f4d".decodeHex()) // Hel
    data.write("5158".decodeHex()) // lo
    serverReaderWithCompression.processNextFrame()

system, Kokoro, schedules our CI jobs by combining a build script with a file
from the `envs` directory that is filled with configuration options:

-   Nightly jobs (Run nightly on the `nightly` branch)
    -   Uses `wheel.sh`, `libtensorflow.sh`, `code_check_full.sh`
-   Continuous jobs (Run on every GitHub commit)
    -   Uses `pycpp.sh`
-   Presubmit jobs (Run on every GitHub PR)
    -   Uses `pycpp.sh`, `code_check_changed_files.sh`

    Otherwise, TeamCity complains "Can't find the previous revision, please commit current settings first".
  - If anything bad happens, feel free to delete the project you created and retry (you may need to apply `Synchronization disabled` before deleting it).  
    
If no errors, your new pipeline will be displayed on TeamCity web UI:

```
Gradle

- To be able to reliably use S3 multipart APIs feature of the SDKs without re-inventing the wheel of pre-signing the each URL in multipart API. This is very tedious to implement with all the scenarios of fault tolerance that's already implemented by the client SDK. The general client SDKs don't support multipart with presigned URLs.

#### Policy management permissions

- admin:CreatePolicy
- admin:DeletePolicy
- admin:GetPolicy
- admin:AttachUserOrGroupPolicy
- admin:ListUserPolicies

#### Heal management permissions

- admin:Heal

#### Service account management permissions

- admin:CreateServiceAccount
- admin:UpdateServiceAccount
- admin:RemoveServiceAccount
- admin:ListServiceAccounts

# TODO(b/304751256): Adjust the numpy pin to a single version, when ready
numpy ~= 1.23.5 ; python_version <= "3.11"
numpy ~= 1.26.0 ; python_version >= "3.12"
wheel ~= 0.41.2
h5py >= 3.10.0
lit ~= 17.0.2
opt_einsum == 3.3.0
astunparse == 1.6.3
dill == 0.3.7
astor == 0.7.1
typing_extensions == 4.8.0
gast == 0.4.0
termcolor == 2.3.0
wrapt == 1.16.0
tblib == 2.0.0

# Install tensorboard, and keras

In previous versions of NodeJS / Browser JavaScript, you would have used "callbacks". Which leads to <a href="http://callbackhell.com/" class="external-link" target="_blank">callback hell</a>.

## Coroutines

!!! warning
    If you are just starting, the tutorial [SQL (Relational) Databases](../tutorial/sql-databases.md){.internal-link target=_blank} that uses SQLAlchemy should be enough.

    Feel free to skip this.

    Peewee is not recommended with FastAPI as it doesn't play well with anything async Python. There are several better alternatives.

!!! info
    These docs assume Pydantic v1.

    Temporarily adopting this point of view (of the *external developer*) can help you feel like it's more obvious where to put the parameters, the Pydantic model for the body, for the response, etc. for that *external API*.

### Create a callback `APIRouter`