amotl

# GitHub Actions Workflow Failures Digest for CrateDB's ecosystem
# https://github.com/crate/crate-clients-tools/issues/148

# Optionally use authentication when running into rate limits.
# It is an (here invalidated) personal access token (classic),
# using the "workflow" scope.

# GITHUB_TOKEN=ghp_00r4G0tGxLDT5RGOwWNw7tZhFnK5fT0uMfoo
# HTTPIE_OPTIONS="--auth-type bearer --auth ${GITHUB_TOKEN}"

amotl

# Connect using SQLAlchemy Core.
import sqlalchemy as sa
from pprint import pp

dburi = "crate://localhost:4200"
query = "SELECT country, mountain, coordinates, height FROM sys.summits ORDER BY country;"


def print_active_servers(connection: sa.engine.Connection):
    print("Active Servers:", connection.connection.dbapi_connection.client.active_servers)

amotl

"""
Attempt to trip a low-memory condition in CrateDB, resulting in a weird error message
like `SQLParseException[ShardCollectContext for 0 already added]`.

This program tries to emulate the MLflow test case `test_search_runs_returns_expected_results_with_large_experiment`,
succeeded by a `DELETE FROM` table truncation operation.

Remark: It did not work out well. This program trips `OutOfMemoryError[Java heap space]`
        right away. Please use the MLflow test case reproducer demonstrated at:
        https://github.com/crate-workbench/mlflow-cratedb/issues/53#issuecomment-1927234463

amotl

"""
## About

Regression with CrateDB nightly-5.7.0-2024-01-26-00-02.
The assembled list of ParameterSymbols is invalid. Missing parameters.

-- https://github.com/crate/crate/issues/15488

## Setup

amotl

-- Demo: SQL "UPDATE ... FROM" for upsert/merge operations.
-- Reflecting Meltano's PostgreSQL data loader (target) adapter.
--
-- Usage:
--
--   psql postgresql://postgres@localhost:5432/ < update-from-merges.sql
--   crash --host http://crate@localhost:4200/ < update-from-merges.sql

-- Make a blank slate.
DROP TABLE IF EXISTS main;

amotl

/**
 * CrateDB: Exercise storing and searching by vectors using its "FLOAT_VECTOR" and "KNN_MATCH".
 * The example uses euclidean distance for vector similarity search.
 *
 * Synopsis::
 *
 *   docker run --rm -it --publish=4200:4200 crate/crate:nightly -Cdiscovery.type=single-node
 *   crash < cratedb-vector-knn-exercise.sql
 *
 * Resources:

amotl

/**
 * PostgreSQL/pgvector: Exercise storing and searching by vectors using its operators.
 * The example uses euclidean distance for vector similarity search.
 *
 * Synopsis::
 *
 *   docker run --rm -it --publish=5432:5432 --env "POSTGRES_HOST_AUTH_METHOD=trust" ankane/pgvector postgres -c log_statement=all
 *   psql postgresql://postgres@localhost < pgvector-vector-knn-exercise.sql
 *
 * Resources:

amotl

"""
pip install requests 'requests-cache<2'
"""

import os

import requests_cache
import typing as t

from langchain.embeddings import OpenAIEmbeddings

amotl

import typing as t

import pytest


def monkeypatch_pytest_notebook_treat_cell_exit_as_notebook_skip():
    """
    Patch `pytest-notebook`, in fact `nbclient.client.NotebookClient`,
    to propagate cell-level `pytest.exit()` invocations as signals
    to mark the whole notebook as skipped.

amotl

import logging
import sys

import requests

from langchain.schema import Document
from langchain.text_splitter import CharacterTextSplitter
from langchain.vectorstores import CrateDBVectorSearch
from langchain.embeddings import OpenAIEmbeddings