rust-data-processing 0.3.3

# THIS FILE IS AUTOMATICALLY GENERATED BY CARGO
#
# When uploading crates to the registry Cargo will automatically
# "normalize" Cargo.toml files for maximal compatibility
# with all versions of Cargo and also rewrite `path` dependencies
# to registry (e.g., crates.io) dependencies.
#
# If you are reading this file be aware that the original Cargo.toml
# will likely look very different (and much more reasonable).
# See Cargo.toml.orig for the original contents.

[package]
edition = "2024"
rust-version = "1.85"
name = "rust-data-processing"
version = "0.3.3"
build = false
exclude = [
    "bindings/",
    "spikes/",
    "Planning/",
    "python-wrapper/",
    "docs/python/",
    "docs/landing/",
    "README.md",
    "test_run.log",
    ".github/",
    "scripts/",
]
autolib = false
autobins = false
autoexamples = false
autotests = false
autobenches = false
description = "Schema-first ingestion (CSV, JSON, Parquet, Excel) into an in-memory DataSet, plus Polars-backed pipelines, SQL, profiling, validation, and map/reduce-style processing."
documentation = "https://docs.rs/rust-data-processing"
readme = "README_CRATE.md"
keywords = [
    "csv",
    "json",
    "parquet",
    "dataset",
    "etl",
]
categories = [
    "database-implementations",
    "parser-implementations",
]
license = "MIT OR Apache-2.0"
repository = "https://github.com/scorpio-datalake/rust-data-processing"
resolver = "2"

[package.metadata.docs.rs]
cargo-args = [
    "-j",
    "1",
]

[features]
arrow = ["dep:arrow"]
ci_expanded = [
    "deep_tests",
    "excel_test_writer",
    "arrow",
    "serde_arrow",
]
cloud_connectors = [
    "object_store",
    "delta_lake",
    "snowflake",
    "file_transfer",
]
db_connectorx = [
    "arrow",
    "dep:connectorx",
    "dep:postgres-protocol",
    "dep:tokio-postgres",
    "connectorx/src_postgres",
    "connectorx/src_mysql",
    "connectorx/src_mssql",
    "connectorx/src_oracle",
]
deep_tests = []
default = ["sql"]
delta_lake = ["object_store"]
excel = ["dep:calamine"]
excel_test_writer = [
    "excel",
    "dep:rust_xlsxwriter",
]
file_transfer = [
    "dep:libssh2-sys",
    "dep:ssh2",
    "dep:suppaftp",
    "dep:url",
    "dep:webpki-roots",
]
integration_full = [
    "db_connectorx",
    "cloud_connectors",
    "excel",
]
kafka = ["dep:rdkafka"]
object_store = [
    "dep:object_store",
    "dep:tokio",
    "dep:url",
    "dep:ureq",
]
serde_arrow = [
    "arrow",
    "dep:serde_arrow",
]
snowflake = ["object_store"]
sql = ["dep:polars-sql"]

[lib]
name = "rust_data_processing"
path = "src/lib.rs"

[[bin]]
name = "generate_people_xlsx_fixture"
path = "src/bin/generate_people_xlsx_fixture.rs"
required-features = ["excel_test_writer"]

[[bin]]
name = "rust-data-processing"
path = "src/main.rs"

[[example]]
name = "file_transfer_ingest"
path = "examples/file_transfer_ingest.rs"
required-features = ["cloud_connectors"]

[[example]]
name = "kafka_elt_byo_load"
path = "examples/kafka_elt_byo_load.rs"
required-features = ["kafka"]

[[example]]
name = "kafka_elt_stream"
path = "examples/kafka_elt_stream.rs"
required-features = ["kafka"]

[[test]]
name = "arrow_record_batches"
path = "tests/arrow_record_batches.rs"

[[test]]
name = "csv_ingestion"
path = "tests/csv_ingestion.rs"

[[test]]
name = "dataframe_centric_pipeline_fixtures"
path = "tests/dataframe_centric_pipeline_fixtures.rs"

[[test]]
name = "deep_tests"
path = "tests/deep_tests.rs"

[[test]]
name = "excel_ingestion"
path = "tests/excel_ingestion.rs"

[[test]]
name = "excel_snippets_fixtures"
path = "tests/excel_snippets_fixtures.rs"

[[test]]
name = "file_transfer_ftp_integration"
path = "tests/file_transfer_ftp_integration.rs"
required-features = ["cloud_connectors"]

[[test]]
name = "ghcn_json_xml_parquet_pipeline_fixtures"
path = "tests/ghcn_json_xml_parquet_pipeline_fixtures.rs"

[[test]]
name = "ingestion_observability"
path = "tests/ingestion_observability.rs"

[[test]]
name = "json_ingestion"
path = "tests/json_ingestion.rs"

[[test]]
name = "json_parquet_excel_snippets_fixtures"
path = "tests/json_parquet_excel_snippets_fixtures.rs"

[[test]]
name = "kafka_elt_load"
path = "tests/kafka_elt_load.rs"
required-features = ["kafka"]

[[test]]
name = "mapping_spec"
path = "tests/mapping_spec.rs"

[[test]]
name = "ordered_batch_ingestion"
path = "tests/ordered_batch_ingestion.rs"

[[test]]
name = "parquet_ingestion"
path = "tests/parquet_ingestion.rs"

[[test]]
name = "parquet_snippets_fixtures"
path = "tests/parquet_snippets_fixtures.rs"

[[test]]
name = "partition_discovery"
path = "tests/partition_discovery.rs"

[[test]]
name = "path_from_directory_scan_fixtures"
path = "tests/path_from_directory_scan_fixtures.rs"

[[test]]
name = "sql"
path = "tests/sql.rs"

[[test]]
name = "student_etl_fixtures"
path = "tests/student_etl_fixtures.rs"

[[test]]
name = "unified_ingestion"
path = "tests/unified_ingestion.rs"

[[test]]
name = "watermark_incremental"
path = "tests/watermark_incremental.rs"

[[bench]]
name = "ingestion"
path = "benches/ingestion.rs"
harness = false
required-features = ["deep_tests"]

[[bench]]
name = "map_reduce"
path = "benches/map_reduce.rs"
harness = false
required-features = ["deep_tests"]

[[bench]]
name = "outliers"
path = "benches/outliers.rs"
harness = false
required-features = ["deep_tests"]

[[bench]]
name = "pipelines"
path = "benches/pipelines.rs"
harness = false
required-features = ["deep_tests"]

[[bench]]
name = "profiling"
path = "benches/profiling.rs"
harness = false
required-features = ["deep_tests"]

[[bench]]
name = "validation"
path = "benches/validation.rs"
harness = false
required-features = ["deep_tests"]

[dependencies.arrow]
version = "54"
optional = true

[dependencies.calamine]
version = "0.33.0"
optional = true

[dependencies.connectorx]
version = "0.4.5"
features = [
    "dst_arrow",
    "fptr",
]
optional = true
default-features = false

[dependencies.csv]
version = "1.3.1"

[dependencies.glob]
version = "0.3"

[dependencies.libssh2-sys]
version = "0.3.1"
features = ["vendored-openssl"]
optional = true

[dependencies.object_store]
version = "0.13.1"
features = [
    "aws",
    "gcp",
    "azure",
    "fs",
]
optional = true
default-features = false

[dependencies.parquet]
version = "57"

[dependencies.polars]
version = "0.53.0"
features = [
    "lazy",
    "csv",
    "parquet",
    "json",
    "fmt",
    "ipc",
]

[dependencies.polars-sql]
version = "0.53.0"
optional = true

[dependencies.postgres-protocol]
version = ">=0.6.12, <0.7"
optional = true

[dependencies.quick-xml]
version = "0.37.5"

[dependencies.rayon]
version = "1"

[dependencies.rdkafka]
version = "0.36.2"
features = ["cmake-build"]
optional = true
default-features = false

[dependencies.rust_xlsxwriter]
version = "0.93"
optional = true

[dependencies.serde]
version = "1"
features = ["derive"]

[dependencies.serde_arrow]
version = "0.14.0"
features = ["arrow-54"]
optional = true
default-features = false

[dependencies.serde_json]
version = "1.0.139"

[dependencies.sha2]
version = "0.10"

[dependencies.ssh2]
version = "0.9.4"
optional = true

[dependencies.suppaftp]
version = "6.3.0"
features = ["rustls"]
optional = true
default-features = false

[dependencies.thiserror]
version = "2.0.11"

[dependencies.tokio]
version = "1"
features = ["rt"]
optional = true

[dependencies.tokio-postgres]
version = ">=0.7.18, <0.8"
features = ["runtime"]
optional = true
default-features = false

[dependencies.ureq]
version = "2.12"
optional = true
default-features = false

[dependencies.url]
version = "2"
optional = true

[dependencies.walkdir]
version = "2"

[dependencies.webpki-roots]
version = "0.26"
optional = true

[dev-dependencies.criterion]
version = "0.5"

[profile.bench]
debug = 0
strip = "debuginfo"

[profile.dev]
debug = 1

[profile.integration]
opt-level = 1
lto = false
codegen-units = 16
debug = 1
inherits = "release"
strip = "debuginfo"

[profile.test]
debug = 0
strip = "debuginfo"