tests: added GroupsParserTests

This commit is contained in:
vantablackTea
2026-09-14 16:33:39 +03:00
parent 02bfe60e8b
commit 8d04c10f48
@@ -0,0 +1,162 @@
package agzam.vyatsu.data.remote
import agzam.vyatsu.domain.model.DateRange
import agzam.vyatsu.domain.model.Season
import org.junit.Assert.assertEquals
import org.junit.Assert.assertTrue
import org.junit.Test
import java.io.InputStream
class GroupsParserTest {
private val faculty = "Факультет автоматики и вычислительной техники (ОРУ)"
private val facultyShort = "ФАВТ (Факультет автоматики и вычислительной техники)"
private fun page(vararg blocks: String): String = blocks.joinToString("\n")
private fun facultyDiv(name: String): String = """<div class="fak_name">$name</div>"""
private fun groupDiv(periodId: String, name: String): String =
"""<div class="grpPeriod" data-grp_period_id="$periodId">$name</div>"""
private fun pdfLink(url: String): String = """<a href="$url">x</a>"""
private fun parse(html: String): ParsedGroups =
GroupsParser.parseStream(html.byteInputStream(Charsets.UTF_8))
private fun parseLineByLine(html: String): ParsedGroups {
val lines = html.split("\n").map { "$it\n" }
var index = 0
val stream = object : InputStream() {
override fun read(): Int = -1
override fun read(b: ByteArray, off: Int, len: Int): Int {
if (index >= lines.size) return -1
val bytes = lines[index].toByteArray(Charsets.UTF_8)
val n = minOf(len, bytes.size)
System.arraycopy(bytes, 0, b, off, n)
index++
return n
}
}
return GroupsParser.parseStream(stream)
}
@Test
fun dropsLastDigitFromPeriodId() {
val html = page(
facultyDiv(faculty),
groupDiv("123456", "ИВТб-3304-05-00 (бакалавриат)")
)
val group = parse(html).groups.single()
assertEquals("12345", group.id)
assertEquals("ИВТб-3304-05-00", group.name)
assertEquals(facultyShort, group.faculty)
}
@Test
fun removesParenthesizedSuffixFromGroupName() {
val html = page(facultyDiv(faculty), groupDiv("777", "Группа А (подгруппа)"))
val group = parse(html).groups.single()
assertEquals("77", group.id)
assertEquals("Группа А", group.name)
}
@Test
fun keepsFacultyNameWhenNoShorthand() {
val html = page(facultyDiv("Некий институт (ОРУ)"), groupDiv("5555", "Гр"))
val group = parse(html).groups.single()
assertEquals("Некий институт", group.faculty)
}
@Test
fun facultyAppliesToGroupsUntilChanged() {
val html = page(
facultyDiv(faculty),
groupDiv("1111", "А-1"),
groupDiv("2222", "А-2"),
facultyDiv("Юридический институт (факультет) (ОРУ)"),
groupDiv("3333", "Ю-1")
)
val faculties = parse(html).groups.map { it.faculty }
assertEquals(
listOf(facultyShort, facultyShort, "ЮИ (Юридический институт)"),
faculties
)
}
@Test
fun parsesDateRangesGroupedBySeason() {
val html = page(
pdfLink("/reports/schedule/Group/12345_1_14092026_27092026.pdf"),
pdfLink("/reports/schedule/Group/12345_1_28092026_11102026.pdf"),
pdfLink("/reports/schedule/Group/12345_2_09022027_22022027.pdf")
)
val ranges = parse(html).dateRanges
assertEquals(
listOf(DateRange("14092026", "27092026"), DateRange("28092026", "11102026")),
ranges["12345"]?.get(Season.AUTUMN)
)
assertEquals(
listOf(DateRange("09022027", "22022027")),
ranges["12345"]?.get(Season.SPRING)
)
}
@Test
fun ignoresUnrelatedLinks() {
val html = page(
pdfLink("/other/12345_1_14092026_27092026.pdf"),
pdfLink("/reports/schedule/Group/123_1_14092026_27092026.pdf")
)
assertTrue(parse(html).dateRanges.isEmpty())
}
@Test
fun extractsRangesFromTextBeforeTag() {
val html = "/reports/schedule/Group/12345_1_14092026_27092026.pdf " + pdfLink("#")
val ranges = parse(html).dateRanges
assertEquals(
listOf(DateRange("14092026", "27092026")),
ranges["12345"]?.get(Season.AUTUMN)
)
}
@Test
fun greaterThanInsideQuotedAttributeDoesNotBreakTag() {
val html = page(
facultyDiv(faculty),
"""<div class="grpPeriod" data-grp_period_id="555" title="a>b">Группа Б</div>"""
)
val group = parse(html).groups.single()
assertEquals("55", group.id)
assertEquals("Группа Б", group.name)
}
@Test
fun multipleReadsProduceSameResultAsSingleRead() {
val html = page(
facultyDiv(faculty),
groupDiv("123456", "ИВТб-3304-05-00 (бакалавриат)"),
pdfLink("/reports/schedule/Group/12345_1_14092026_27092026.pdf"),
groupDiv("7654321", "ИВТб-3304-05-01"),
pdfLink("/reports/schedule/Group/765432_2_09022027_22022027.pdf")
)
assertEquals(parse(html), parseLineByLine(html))
}
}