Crates.io | rinfluxdb-polars |
lib.rs | rinfluxdb-polars |
version | 0.2.0 |
source | src |
created_at | 2021-11-27 20:24:56.258901 |
updated_at | 2021-11-27 20:24:56.258901 |
description | A library for querying and posting data to InfluxDB |
homepage | |
repository | https://gitlab.com/claudiomattera/rinfluxdb |
max_upload_size | |
id | 488548 |
size | 22,118 |
A library for querying and sending data to InfluxDB.
https://gitlab.com/claudiomattera/rinfluxdb
Serialize data to InfluxDB line protocol;
Build InfluxQL queries in Rust;
Build FLUX queries in Rust;
Parse responses from InfluxDB;
(Optional) client based on Reqwest to perform common queries.
Parse dataframes from InfluxDB JSON returned for InfluxQL queries;
Parse dataframes from InfluxDB annotated CSV returned for FLUX queries;
(Optional) wrapper around Reqwest objects to construct requests and parse responses;
The line protocol is used to send data to InfluxDB.
use rinfluxdb::line_protocol::LineBuilder;
use chrono::{TimeZone, Utc};
let line = LineBuilder::new("location")
.insert_field("latitude", 55.383333)
.insert_field("longitude", 10.383333)
.insert_tag("city", "Odense")
.set_timestamp(Utc.ymd(2014, 7, 8).and_hms(9, 10, 11))
.build();
assert_eq!(line.measurement(), &"location".into());
assert_eq!(line.field("latitude"), Some(&55.383333.into()));
assert_eq!(line.field("longitude"), Some(&10.383333.into()));
assert_eq!(line.tag("city"), Some(&"Odense".into()));
assert_eq!(line.timestamp(), Some(&Utc.ymd(2014, 7, 8).and_hms(9, 10, 11)));
assert_eq!(
line.to_string(),
"location,city=Odense latitude=55.383333,longitude=10.383333 1404810611000000000"
);
InfluxQL queries can be built using influxql::QueryBuilder
.
use rinfluxdb::influxql::QueryBuilder;
use chrono::{TimeZone, Utc};
let query = QueryBuilder::from("indoor_environment")
.field("temperature")
.field("humidity")
.start(Utc.ymd(2021, 3, 7).and_hms(21, 0, 0))
.build();
assert_eq!(
query.as_ref(),
"SELECT temperature, humidity \
FROM indoor_environment \
WHERE time > '2021-03-07T21:00:00Z'",
);
FLUX queries can be built using flux::QueryBuilder
.
use rinfluxdb::types::Duration;
use rinfluxdb::flux::QueryBuilder;
let query = QueryBuilder::from("telegraf/autogen")
.range_start(Duration::Minutes(-15))
.filter(
r#"r._measurement == "cpu" and
r._field == "usage_system" and
r.cpu == "cpu-total""#
)
.build();
assert_eq!(
query.as_ref(),
r#"from(bucket: "telegraf/autogen")
|> range(start: -15m)
|> filter(fn: (r) =>
r._measurement == "cpu" and
r._field == "usage_system" and
r.cpu == "cpu-total"
)
|> yield()"#,
);
When sending a query to InfluxDB, it will reply with either JSON or annotated CSV content containing a list of dataframes. This library allows to parse such replies to user-defined dataframe types.
A dataframe must be constructable from its name (a string), its index (a vector of instants) and its columns (a mapping of column names to vector of values).
A dataframe implementation only needs to implement this trait to be used with this crate.
I. e., as long as trait TryFrom<(String, Vec<DateTime<Utc>>, HashMap<String, Vec<Value>>), Error = E>
is implemented for a given type DF
(and type E
implements Into<ParseError>
), the parser can use it to construct the final objects.
A dummy implementation of a dataframe is available as dataframe::DataFrame
, but the trait can be implemented for many other existing libraries.
JSON responses to InfluxQL queries can be parsed to dataframes.
use rinfluxdb::influxql::{ResponseError, StatementResult, from_str};
use rinfluxdb::dataframe::DataFrame;
let input: String = todo!();
let response: Result<Vec<StatementResult<DataFrame>>, ResponseError> =
from_str(&input);
Annotated CSV responses to FLUX queries can be parsed.
use rinfluxdb::flux::{ResponseError, from_str};
use rinfluxdb::dataframe::DataFrame;
let input: String = todo!();
let response: Result<DataFrame, ResponseError> = from_str(&input);
The functions shown above can be used to serialize and deserialize queries and data to raw text, and they can be integrated into existing applications. In alternative, this library also implements an optional client API based on Reqwest to directly interact with an InfluxDB instance. Both blocking and asynchronous clients are available, and they support common queries.
Clients are enabled using the client
Cargo feature.
# use std::collections::HashMap;
# use url::Url;
#
use rinfluxdb::influxql::QueryBuilder;
use rinfluxdb::influxql::blocking::Client;
use rinfluxdb::dataframe::DataFrame;
let client = Client::new(
Url::parse("https://example.com/")?,
Some(("username", "password")),
)?;
let query = QueryBuilder::from("indoor_environment")
.database("house")
.field("temperature")
.field("humidity")
.build();
let dataframe: DataFrame = client.fetch_dataframe(query)?;
println!("{}", dataframe);
let query = QueryBuilder::from("indoor_environment")
.database("house")
.field("temperature")
.field("humidity")
.group_by("room")
.build();
let tagged_dataframes: HashMap<String, DataFrame> =
client.fetch_dataframes_by_tag(query, "room")?;
for (tag, dataframe) in tagged_dataframes {
println!("{}: {}", tag, dataframe);
}
# Ok::<(), rinfluxdb::influxql::ClientError>(())
unimplemented!()
# use url::Url;
#
use rinfluxdb::line_protocol::LineBuilder;
use rinfluxdb::line_protocol::blocking::Client;
let client = Client::new(
Url::parse("https://example.com/")?,
Some(("username", "password")),
)?;
let lines = vec![
LineBuilder::new("measurement")
.insert_field("field", 42.0)
.build(),
LineBuilder::new("measurement")
.insert_field("field", 43.0)
.insert_tag("tag", "value")
.build(),
];
client.send("database", &lines)?;
# Ok::<(), rinfluxdb::line_protocol::ClientError>(())
This crate communicates with InfluxDB instances over HTTP(s). The content of HTTP requests and responses must follow InfluxDB conventions and protocols, but they can otherwise be customized, e.g. by adding basic authentication.
In order to ensure maximal freedom, a tiny wrapper is constructed around Reqwest's Client
, so that it can create a request builder already prepared to communicate with InfluxDB.
Such builder can be then converted to a regular request builder and executed.
# use url::Url;
#
use rinfluxdb::influxql::Query;
// Bring into scope the trait implementation
use rinfluxdb::influxql::blocking::InfluxqlClientWrapper;
// Create Reqwest client
let client = reqwest::blocking::Client::new();
// Create InfluxQL request
let base_url = Url::parse("https://example.com")?;
let mut builder = client
// (this is a function added by the trait above)
.influxql(&base_url)?
// (this functions are defined on influxql::RequestBuilder)
.database("house")
.query(Query::new("SELECT temperature FROM indoor_temperature"))
// (this function returns a regular Reqwest builder)
.into_reqwest_builder();
// Now this is a regular Reqwest builder, and can be customized as usual
if let Some((username, password)) = Some(("username", "password")) {
builder = builder.basic_auth(username, Some(password));
}
// Create a request from the builder
let request = builder.build()?;
// Execute the request through Reqwest and obtain a response
let response = client.execute(request)?;
# Ok::<(), rinfluxdb::influxql::ClientError>(())
Similarly, a tiny wrapper is constructed around Reqwest's Response
, so that a new function is added to parse dataframes from it.
# use std::collections::HashMap;
#
# use url::Url;
#
use rinfluxdb::influxql::Query;
use rinfluxdb::influxql::StatementResult;
use rinfluxdb::influxql::blocking::InfluxqlClientWrapper;
use rinfluxdb::dataframe::DataFrame;
// Bring into scope the trait implementation
use rinfluxdb::influxql::blocking::InfluxqlResponseWrapper;
// Create Reqwest client
let client = reqwest::blocking::Client::new();
// Create InfluxQL request
let base_url = Url::parse("https://example.com")?;
let mut request = client
.influxql(&base_url)?
.database("house")
.query(Query::new("SELECT temperature FROM indoor_temperature"))
.into_reqwest_builder()
.build()?;
// Execute the request through Reqwest and obtain a response
let response = client.execute(request)?;
// Return an error if response status is not 200
// (this is a function from Reqwest's response)
let response = response.error_for_status()?;
// Parse the response from JSON to a list of dataframes
// (this is a function added by the trait above)
let results: Vec<StatementResult<DataFrame>> = response.dataframes()?;
# Ok::<(), rinfluxdb::influxql::ClientError>(())
Wrappers are defined for both Reqwest's blocking API (influxql::blocking::InfluxqlClientWrapper
, influxql::blocking::InfluxqlResponseWrapper
) and Reqwest's asynchronous API (influxql::r#async::InfluxqlClientWrapper
, influxql::r#async::InfluxqlResponseWrapper
), and are enabled using the client
Cargo feature.
This crate is a simple aggregator over smaller crates, each enabled by a Cargo feature and implementing a specific part of InfluxDB support.
rinfluxdb
├── rinfluxdb-types
├── rinfluxdb-lineprotocol
├── rinfluxdb-influxql
├── rinfluxdb-flux
└── rinfluxdb-dataframe
Clients can either depend on rinfluxdb
enabling the necessary features, or they can depend explicitly on the rinfluxdb-*
crates.
[dependencies.rinfluxdb]
version = "0.2.0"
features = ["lineprotocol", "influxql", "client"]
# Or
[dependencies]
rinfluxdb-lineprotocol = { version = "0.2.0", features = ["client"] }
rinfluxdb-influxql = { version = "0.2.0", features = ["client"] }
This crate supports the following Cargo features.
lineprotocol
: re-exports rinfluxdb-lineprotocol
crate;influxql
: re-exports rinfluxdb-influxql
crate;flux
: re-exports rinfluxdb-flux
crate;dataframe
: re-exports rinfluxdb-dataframe
crate;client
: enables feature client
in all rinfluxdb-*
crates.When feature client
is enabled, the crates define clients for line protocol, InfluxQL and Flux.
Clients are implemented using Reqwest, and are available both for blocking and async mode.
Copyright Claudio Mattera 2021
You are free to copy, modify, and distribute this application with attribution under the terms of either
at your option.
This project is entirely original work, and it is not affiliated with nor endorsed in any way by InfluxData.