Commit 7dd00f76 authored by Wohlgemuth, Jason's avatar Wohlgemuth, Jason
Browse files

feat: Remove more TODO comments; Add support for repeatable fields in BagInfo

parent f561bb2c

acorn-lib/src/io/api/mod.rs

+0 −1
Original line number Diff line number Diff line
@@ -684,7 +684,6 @@ where 
        | Err(e) => Err(format!("Failed to parse ORCiD search response: {e}")),

    }

}

// TODO: Support OR statements for field values (e.g., affiliation-org-name:("University of Plymouth" OR "Plymouth University"))

/// Construct a query string for an endpoint API query from a list of field-value pairs, a list of fields, and a list of fields with boosted relevance.

///

/// The query string is constructed by joining the following parts with "&":

acorn-lib/src/io/bagit.rs

+21 −16
Original line number Diff line number Diff line
@@ -81,26 +81,26 @@ pub struct Bag { 
#[builder(start_fn = init)]

pub struct BagInfo {

    /// Organization transferring the content (Source-Organization)

    pub organization: Option<String>,

    pub organization: Option<Vec<String>>,

    /// Mailing address of the source organization (Organization-Address)

    pub organization_address: Option<String>,

    pub organization_address: Option<Vec<String>>,

    /// Person responsible for the content transfer (Contact-Name)

    pub contact_name: Option<String>,

    pub contact_name: Option<Vec<String>>,

    /// International format telephone number of responsible person (Contact-Phone)

    pub contact_phone: Option<String>,

    pub contact_phone: Option<Vec<String>>,

    /// Fully qualified email address of responsible person (Contact-Email)

    pub contact_email: Option<String>,

    pub contact_email: Option<Vec<String>>,

    /// Brief explanation of contents and provenance (External-Description)

    pub description: Option<String>,

    pub description: Option<Vec<String>>,

    /// Date (YYYY-MM-DD) that content was prepared for transfer (Bagging-Date)

    pub date: Option<String>,

    /// Sender-supplied identifier for the bag (External-Identifier)

    pub identifier: Option<String>,

    pub identifier: Option<Vec<String>>,

    /// Size or approximate size of the bag with abbreviation (Bag-Size)

    pub size: Option<String>,

    /// Bag count as (N, T) tuple where N is ordinal position and T is total (Bag-Count)

    /// Bag count values as (N, T) tuples where N is ordinal position and T is total

    /// T is None if unknown ("?")

    pub count: Option<(u32, Option<u32>)>,

    pub count: Option<Vec<(u32, Option<u32>)>>,

}

impl Bag {

    /// Verify the completeness and correctness of a bag
@@ -280,25 +280,30 @@ impl BagInfo { 
            size,

            count,

        } = self;

        let base = [

        let repeatable = [

            ("Source-Organization", organization),

            ("Organization-Address", organization_address),

            ("Contact-Name", contact_name),

            ("Contact-Phone", contact_phone),

            ("Contact-Email", contact_email),

            ("External-Description", description),

            ("Bagging-Date", date),

            ("External-Identifier", identifier),

            ("Bag-Size", size),

        ]

        .into_iter()

        .filter_map(|(key, value)| value.as_ref().map(|v| (key.to_string(), v.clone())));

        // Option as iterator: Some(item) -> one element, None -> empty.

        let count = count.iter().map(|(index, total)| {

        .flat_map(|(key, values)| {

            values

                .iter()

                .flat_map(|items| items.iter())

                .map(move |value| (key.to_string(), value.clone()))

        });

        let count = count.iter().flat_map(|items| items.iter()).map(|(index, total)| {

            let total = total.map_or_else(|| "?".to_string(), |value| value.to_string());

            ("Bag-Count".to_string(), format!("{index} of {total}"))

        });

        base.chain(count).collect()

        let single = [("Bagging-Date", date.clone()), ("Bag-Size", size.clone())]

            .into_iter()

            .filter_map(|(key, value)| value.map(|v| (key.to_string(), v)));

        repeatable.chain(count).chain(single).collect()

    }

}

impl Default for BagInfo {

acorn-lib/src/io/tests/mod.rs

+51 −3
Original line number Diff line number Diff line
@@ -7,7 +7,9 @@ use crate::io::{ 
    filter_ignored, image_paths, InputOutput,

};

use crate::{Location, Repository, Scheme};

use std::fs::{create_dir_all, read_to_string, remove_dir_all};

use std::path::PathBuf;

use std::time::{SystemTime, UNIX_EPOCH};



const ENDPOINTS: &str = r#"{

    "endpoints": [
@@ -75,6 +77,7 @@ const ENDPOINTS: &str = r#"{ 
        }

    ]

}"#;

const TEST_BAGGING_DATE: &str = "2026-02-22";



fn fixtures_dir() -> PathBuf {

    PathBuf::from(env!("CARGO_MANIFEST_DIR")).parent().unwrap().join("tests/fixtures")
@@ -219,7 +222,6 @@ fn test_runners() { 
    let config = ApplicationConfiguration::read_json(path).unwrap();

    insta::assert_snapshot!("acorn__io__config__tests__runners", format!("{:#?}", config.runners));

}



#[test]

fn test_archive() {

    let path = fixtures_dir().join("data");
@@ -229,8 +231,8 @@ fn test_archive() { 
#[test]

fn test_bagit() {

    let info = BagInfo::init()

        .organization("Oak Ridge National Laboratory".to_string())

        .contact_name("Jason Wohlgemuth".to_string())

        .organization(vec!["Oak Ridge National Laboratory".to_string()])

        .contact_name(vec!["Jason Wohlgemuth".to_string()])

        .build();

    let bag = Bag::init()

        .base_directory(fixtures_dir().join("data").display().to_string())
@@ -247,6 +249,52 @@ fn test_bagit() { 
    }

}

#[test]

fn test_baginfo_save_repeatable_fields() {

    let stamp = SystemTime::now().duration_since(UNIX_EPOCH).unwrap().as_nanos();

    let dir = std::env::temp_dir().join(format!("acorn-baginfo-{stamp}"));

    create_dir_all(dir.clone()).unwrap();

    let info = BagInfo::init()

        .organization(vec!["Org One".to_string(), "Org Two".to_string()])

        .contact_name(vec!["Alice".to_string(), "Bob".to_string()])

        .count(vec![(1, Some(3)), (2, None)])

        .date(TEST_BAGGING_DATE.to_string())

        .size("1 MB".to_string())

        .build();

    let path = info.save(dir.clone()).unwrap();

    assert_eq!(path, dir);

    let content = read_to_string(dir.join("bag-info.txt")).unwrap();

    let lines = content.lines().collect::<Vec<_>>();

    let expected_date = format!("Bagging-Date: {TEST_BAGGING_DATE}");

    assert_eq!(

        lines,

        vec![

            "Source-Organization: Org One",

            "Source-Organization: Org Two",

            "Contact-Name: Alice",

            "Contact-Name: Bob",

            "Bag-Count: 1 of 3",

            "Bag-Count: 2 of ?",

            expected_date.as_str(),

            "Bag-Size: 1 MB",

        ]

    );

    remove_dir_all(dir).unwrap();

}

#[test]

fn test_baginfo_save_single_value_fields_only() {

    let stamp = SystemTime::now().duration_since(UNIX_EPOCH).unwrap().as_nanos();

    let dir = std::env::temp_dir().join(format!("acorn-baginfo-single-{stamp}"));

    create_dir_all(dir.clone()).unwrap();

    let info = BagInfo::init().date(TEST_BAGGING_DATE.to_string()).size("42 KB".to_string()).build();

    let path = info.save(dir.clone()).unwrap();

    assert_eq!(path, dir);

    let content = read_to_string(dir.join("bag-info.txt")).unwrap();

    let lines = content.lines().collect::<Vec<_>>();

    let expected_date = format!("Bagging-Date: {TEST_BAGGING_DATE}");

    assert_eq!(lines, vec![expected_date.as_str(), "Bag-Size: 42 KB"]);

    remove_dir_all(dir).unwrap();

}

#[test]

#[ignore = "Verification is broke on macbook for some reason"]

fn test_bagit_verify() {

    let path = fixtures_dir().join("bag");

acorn-lib/src/powerpoint/mod.rs

+1 −1
Original line number Diff line number Diff line
@@ -162,10 +162,10 @@ pub fn interpolate_values(path: PathBuf, data: ResearchActivity) { 
                String(&'static str, String),

                Bullets(&'static str, Vec<String>),

            }

            // TODO: Use CiteAs API to get in Chicago format

            let replacements = vec![

                Replacement::String("caption", caption),

                Replacement::String("challenge", challenge),

                // TODO: Use CiteAs API to get in Chicago format

                Replacement::String("citation", doi.unwrap_or_else(|| vec!["".to_string()])[0].clone()),

                Replacement::String("email", email),

                Replacement::String("first", first),

acorn-lib/src/schema/mod.rs

+1 −1
Original line number Diff line number Diff line
@@ -708,7 +708,7 @@ impl ContactPoint { 
    #[cfg(feature = "std")]

    pub fn format_with(self, _context: Option<PathBuf>) -> Self {

        // TODO: Resolve ORCiD identifier from first, last, and email

        self

        self.format()

    }

}

impl Default for ContactPoint {