martin
/
PbDbFixer
1
1
Fork 0
Code Issues Pull Requests Projects Releases 14 Wiki Activity

Handle multiple authors in epub3.

This commit is contained in:
Martin Brodbeck 2021-01-29 13:51:36 +01:00
parent 049cd06b65
commit bc2d4730f0
1 changed files with 25 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
src/main.rs
View File

@ -38,8 +38,9 @@ fn get_attribute_file_as(opf: ZipFile) -> Option<String> {
.create_reader(opf); .create_reader(opf);
let mut refines_found = false; let mut refines_found = false;
let mut refines_entries = Vec::new();
let mut is_epub3 = false; let mut is_epub3 = false;
let mut creator_id = String::new(); let mut creator_ids = Vec::new();
for e in parser { for e in parser {
match e { match e {
@ -63,7 +64,7 @@ fn get_attribute_file_as(opf: ZipFile) -> Option<String> {
return Some(attr.value); return Some(attr.value);
} }
if is_epub3 && attr.name.local_name == "id" { if is_epub3 && attr.name.local_name == "id" {
creator_id = attr.value; creator_ids.push("#".to_owned() + attr.value.as_str());
} }
} }
} }
@ -71,9 +72,7 @@ fn get_attribute_file_as(opf: ZipFile) -> Option<String> {
name, attributes, .. name, attributes, ..
}) if name.local_name == "meta" => { }) if name.local_name == "meta" => {
if attributes.iter().any(|attr| { if attributes.iter().any(|attr| {
attr.name.local_name == "refines" attr.name.local_name == "refines" && creator_ids.contains(&attr.value)
&& (attr.value == creator_id
|| attr.value == "#".to_owned() + creator_id.as_str())
}) && attributes }) && attributes
.iter() .iter()
.any(|attr| attr.name.local_name == "property" && attr.value == "file-as") .any(|attr| attr.name.local_name == "property" && attr.value == "file-as")
@ -83,7 +82,8 @@ fn get_attribute_file_as(opf: ZipFile) -> Option<String> {
} }
Ok(XmlEvent::Characters(value)) => { Ok(XmlEvent::Characters(value)) => {
if refines_found == true { if refines_found == true {
return Some(value); refines_entries.push(value);
refines_found = false;
} }
} }
Ok(XmlEvent::StartElement { .. }) => { Ok(XmlEvent::StartElement { .. }) => {
@ -98,6 +98,12 @@ fn get_attribute_file_as(opf: ZipFile) -> Option<String> {
} }
} }
if refines_entries.len() == 1 {
return Some(refines_entries.remove(0));
} else if refines_entries.len() >= 2 {
return Some(refines_entries.join(" & "));
}
None None
} }
@ -112,6 +118,8 @@ fn main() {
let mut conn = Connection::open("/mnt/ext1/system/explorer-3/explorer-3.db").unwrap(); let mut conn = Connection::open("/mnt/ext1/system/explorer-3/explorer-3.db").unwrap();
let tx = conn.transaction().unwrap(); let tx = conn.transaction().unwrap();
{ {
// Get book ids from entries where we have something like "firstname lastname" in author
// but no "lastname, firstname" in fistauthor
let mut stmt = tx.prepare("SELECT id FROM books_impl WHERE ext LIKE 'epub' AND author LIKE '% %' AND (firstauthor NOT LIKE '%\\,%' ESCAPE '\\' OR firstauthor LIKE '%&amp;%')").unwrap(); let mut stmt = tx.prepare("SELECT id FROM books_impl WHERE ext LIKE 'epub' AND author LIKE '% %' AND (firstauthor NOT LIKE '%\\,%' ESCAPE '\\' OR firstauthor LIKE '%&amp;%')").unwrap();
let mut rows = stmt.query(NO_PARAMS).unwrap(); let mut rows = stmt.query(NO_PARAMS).unwrap();
let mut book_ids: Vec<i32> = Vec::new(); let mut book_ids: Vec<i32> = Vec::new();
@ -119,6 +127,17 @@ fn main() {
book_ids.push(row.get(0).unwrap()); book_ids.push(row.get(0).unwrap());
} }
// Get also book ids from the special case where we have multiple authors (separated by ", " in authors)
// but no ampersand ("&") in firstauthor
let mut stmt = tx.prepare("SELECT id FROM books_impl WHERE ext LIKE 'epub' AND author LIKE '%\\, %' ESCAPE '\\' AND firstauthor NOT LIKE '%&%'").unwrap();
let mut rows = stmt.query(NO_PARAMS).unwrap();
while let Some(row) = rows.next().unwrap() {
book_ids.push(row.get(0).unwrap());
}
book_ids.sort();
book_ids.dedup();
let mut bookentries = Vec::new(); let mut bookentries = Vec::new();
for book_id in book_ids { for book_id in book_ids {