This client is inspired by sckott/habanero.
Crossref
- Crossref search API. The Crossref
crate provides methods matching Crossref API routes:
works
-/works
routemembers
-/members
routeprefixes
-/prefixes
routefunders
-/funders
routejournals
-/journals
routetypes
-/types
routeagency
-/works/{doi}/agency
get DOI minting agency
let client = Crossref::builder().build()?;
If you have an Authorization token for Crossref's Plus service:
let client = Crossref::builder()
.token("token")
.build()?;
Encouraged to use the The Polite Pool:
Good manners = more reliable service
To get into Crossref's polite pool include a email address
let client = Crossref::builder()
.polite("polite@example.com")
.token("your token")
.build()?;
Not all components support queries and there are custom available parameters for each route that supports querying.
For each resource components that supports querying there exist a Query struct: WorksQuery
, MembersQuery
, FundersQuery
. The WorksQuery
also differs from the others by supporting deep paging with cursors and field queries.
otherwise creating queries works the same for all resource components:
let query = WorksQuery::new("Machine Learning")
// field queries supported for `Works`
.field_query(FieldQuery::author("Some Author"))
// filters are specific for each resource component
.filter(WorksFilter::HasOrcid)
.order(Order::Asc)
.sort(Sort::Score);
See this table for a detailed overview of the major components.
There are 3 different targets:
- standalone resource components:
/works
,/members
,funders
,prefixes
,types
that return a list list of the corresponding items and can be specified with queries - Resource component with identifiers:
/works/{doi}?<query>
,/members/{member_id}?<query>
, etc. that returns a single item if found. - combined with the
works
route: The works component can be appended to other resources:/members/{member_id}/works?<query>
etc. that returns a list of matchingWork
items asWorkList
.
This resembles in the enums of the resource components, eg. for Members
:
pub enum Members {
/// target a specific member at `/members/{id}`
Identifier(String),
/// target all members that match the query at `/members?query...`
Query(MembersQuery),
/// target a `Work` for a specific member at `/members/{id}/works?query..`
Works(WorksIdentQuery),
}
All options are supported by the client:
Query Single Item by DOI or ID
Analogous methods exist for all resource components
let work = client.work("10.1037/0003-066X.59.1.29")?;
let agency = client.work_agency("10.1037/0003-066X.59.1.29")?;
let funder = client.funder("funder_id")?;
let member = client.member("member_id")?;
Query
let query = WorksQuery::new("Machine Learning");
// one page of the matching results
let works = client.works(query)?;
Alternatively insert a free form query term directly
let works = client.works("Machine Learning")?;
Combining Routes with the Works
route
For each resource component other than Works
there exist methods to append a WorksQuery
with the ID option /members/{member_id}/works?<query>?
use crossref::*;
fn run() -> Result<()> {
let client = Crossref::builder().build()?;
let works = client.member_works(WorksQuery::new("machine learning")
.sort(Sort::Score).into_ident("member_id"))?;
Ok(())
}
This would be the same as using the [Crossref::works
] method by supplying the combined type
use crossref::*;
fn run() -> Result<()> {
let client = Crossref::builder().build()?;
let works = client.works(WorksQuery::new("machine learning")
.sort(Sort::Score)
.into_combined_query::<Members>("member_id"))?;
Ok(())
}
** Deep paging for Works
**
Deep paging results
Deep paging is supported for all queries, that return a list of Work
, WorkList
.
This function returns a new iterator over pages of Work
, which is returned as bulk of items as a WorkList
by crossref.
Usually a single page WorkList
contains 20 items.
Example
Iterate over all Works
linked to search term Machine Learning
use crossref::{Crossref, WorksQuery, Work};
fn run() -> Result<(), crossref::Error> {
let client = Crossref::builder().build()?;
let all_works: Vec<Work> = client.deep_page(WorksQuery::new("Machine Learning")).flat_map(|x|x.items).collect();
Ok(())
}
Which can be simplified to
use crossref::{Crossref, WorksQuery, Work};
fn run() -> Result<(), crossref::Error> {
let client = Crossref::builder().build()?;
let all_works: Vec<Work> = client.deep_page("Machine Learning").into_work_iter().collect();
Ok(())
}
Iterate over all the pages (WorkList
) of the funder with id funder id
by using a combined query.
A single WorkList
usually holds 20 Work
items.
use crossref::{Crossref, Funders, WorksQuery, Work, WorkList};
fn run() -> Result<(), crossref::Error> {
let client = Crossref::builder().build()?;
let all_funder_work_list: Vec<WorkList> = client.deep_page(WorksQuery::default()
.into_combined_query::<Funders>("funder id")
)
.collect();
Ok(())
}
Iterate over all Work
items of a specfic funder directly.
use crossref::{Crossref, Funders, WorksQuery, Work, WorkList};
fn run() -> Result<(), crossref::Error> {
let client = Crossref::builder().build()?;
let all_works: Vec<Work> = client.deep_page(WorksQuery::default()
.into_combined_query::<Funders>("funder id"))
.into_work_iter()
.collect();
Ok(())
}
cargo install crossref --features cli
Top level subcommands
USAGE:
crossref <SUBCOMMAND>
FLAGS:
-h, --help Prints help information
-V, --version Prints version information
SUBCOMMANDS:
funders Query crossref funders
help Prints this message or the help of the given subcommand(s)
journals Query crossref journals
members Query crossref members
prefixes Query crossref prefixes
types Query crossref types
works Query crossref works
Additional options for the component subcommands (querying, sorting, ordering and limiting is only supported for subcommands <works|funders|members> and is overridden by a present --id
options)
USAGE:
crossref works [FLAGS] [OPTIONS] [SUBCOMMAND]
FLAGS:
-a, --append if the output file already exists, append instead of overwriting the file
-d, --deep-page Enable deep paging. If a limit is set, then the limit takes priority.
-h, --help Prints help information
-s, --silent do not print anything
-V, --version Prints version information
OPTIONS:
-i, --id <id> The id of component.
-l, --limit <limit> limit the amount of results
--offset <offset> Sets an offset where crossref begins to retrieve items.
--order <order> How to order the results: asc or desc
-o <output> output path where the results shall be stored
--polite <polite> The email to use to get into crossref's polite pool
-q, --query <query_terms>... The free form terms for the query
--sample <sample> Request randoms Elements. Overrides all other options.
--sort <sort> How to sort the results, such as updated, indexed, published, issued
--token <token> The token to use for the crossref client
--user-agent <user_agent> The user agent to use for the crossref client
Retrieve a specific work by a doi
crossref works --id "10.1037/0003-066X.59.1.29"
Save the results as json
crossref works --id "10.1037/0003-066X.59.1.29" -o output.json
Retrieve any other components by their ids
crossref <works|journals|members|prefixes|types> --id "10.1037/0003-066X.59.1.29" -o output.json
Some components support additional filtering
crossref <works|funders|members> --query "A search term such as `Machine learning` for works" --limit 10 --offset 200 --order asc
Get Works
of a specific component, such as a member with the id 98
:
crossref works member 98
<prefix|funder|prefix|type
are also supported in the same way.
By default deep paging is disabled, hence the max amount of results of Works
will be 20 (a single crossref page).
By enabling the --deep-page
flag, all available results will be gathered.
To get in to the polite pool supply your email to the request headers with --polite "polite@example.com"
Licensed under either of these:
- Apache License, Version 2.0, (LICENSE-APACHE or https://www.apache.org/licenses/LICENSE-2.0)
- MIT license (LICENSE-MIT or https://opensource.org/licenses/MIT)