Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion analysis/build.gradle.kts
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,8 @@ repositories {
}

dependencies {
implementation("com.github.MaibornWolff:TreeSitterExcavationSite:0.8.0")
// TEMPORARY: composite-build coordinate. REVERT to "com.github.MaibornWolff:TreeSitterExcavationSite:<tag>" before the DC PR.
implementation("de.maibornwolff.treesitter.excavationsite:treesitter-excavationsite")
Comment thread
christian-spahn-mw marked this conversation as resolved.
implementation("io.github.bonede:tree-sitter:0.26.3")
implementation("io.github.bonede:tree-sitter-typescript:0.23.2")
implementation("io.github.bonede:tree-sitter-php:0.23.11")
Expand Down
4 changes: 4 additions & 0 deletions analysis/settings.gradle.kts
Original file line number Diff line number Diff line change
Expand Up @@ -2,3 +2,7 @@ plugins {
id("org.gradle.toolchains.foojay-resolver-convention") version "1.0.0"
}
rootProject.name = "dependacharta-analysis"

// TEMPORARY: composite-build wiring to local TSE for the Python migration iteration phase.
// REVERT before opening the DC PR — see plans/add-python-dependency-support.md Task 7.
includeBuild("../../TreeSitterExcavationSite")
Comment thread
christian-spahn-mw marked this conversation as resolved.
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,8 @@ fun DeclarationType.toNodeType(): NodeType {
DeclarationType.INTERFACE -> NodeType.INTERFACE
DeclarationType.ENUM -> NodeType.ENUM
DeclarationType.ANNOTATION -> NodeType.ANNOTATION
DeclarationType.FUNCTION -> NodeType.FUNCTION
DeclarationType.VARIABLE -> NodeType.VARIABLE
DeclarationType.UNKNOWN -> NodeType.UNKNOWN
}
}
Original file line number Diff line number Diff line change
@@ -1,193 +1,150 @@
package de.maibornwolff.dependacharta.pipeline.analysis.analyzers.python

import de.maibornwolff.dependacharta.pipeline.analysis.analyzers.LanguageAnalyzer
import de.maibornwolff.dependacharta.pipeline.analysis.analyzers.common.utils.nodeAsString
import de.maibornwolff.dependacharta.pipeline.analysis.analyzers.common.utils.withoutFileSuffix
import de.maibornwolff.dependacharta.pipeline.analysis.analyzers.python.queries.PythonDefinitionsQuery
import de.maibornwolff.dependacharta.pipeline.analysis.analyzers.python.queries.PythonImportQuery
import de.maibornwolff.dependacharta.pipeline.analysis.analyzers.python.queries.PythonTypeAttributeQuery
import de.maibornwolff.dependacharta.pipeline.analysis.analyzers.python.queries.PythonTypeIdentifierQuery
import de.maibornwolff.dependacharta.pipeline.analysis.model.*
import de.maibornwolff.dependacharta.pipeline.analysis.analyzers.toNodeType
import de.maibornwolff.dependacharta.pipeline.analysis.model.Dependency
import de.maibornwolff.dependacharta.pipeline.analysis.model.FileInfo
import de.maibornwolff.dependacharta.pipeline.analysis.model.FileReport
import de.maibornwolff.dependacharta.pipeline.analysis.model.Node
import de.maibornwolff.dependacharta.pipeline.analysis.model.NodeType
import de.maibornwolff.dependacharta.pipeline.analysis.model.Path
import de.maibornwolff.dependacharta.pipeline.analysis.model.Type
import de.maibornwolff.dependacharta.pipeline.shared.SupportedLanguage
import org.treesitter.TSNode
import org.treesitter.TSParser
import org.treesitter.TreeSitterPython
import de.maibornwolff.treesitter.excavationsite.api.Declaration
import de.maibornwolff.treesitter.excavationsite.api.DeclarationType
import de.maibornwolff.treesitter.excavationsite.api.ImportDeclaration
import de.maibornwolff.treesitter.excavationsite.api.ImportKind
import de.maibornwolff.treesitter.excavationsite.api.Language
import de.maibornwolff.treesitter.excavationsite.api.TreeSitterDependencies
import de.maibornwolff.treesitter.excavationsite.api.UsedType

class PythonAnalyzer(
private val fileInfo: FileInfo
) : LanguageAnalyzer {
private val python = TreeSitterPython()
private val definitionsQuery = PythonDefinitionsQuery(python)
private val importQuery = PythonImportQuery(python)
private val identifierQuery = PythonTypeIdentifierQuery(python)
private val attributeQuery = PythonTypeAttributeQuery(python)
private companion object {
const val INIT = "__init__"
const val PY_SUFFIX = "py"
const val SEPARATOR = "."
}

override fun analyze(): FileReport {
val rootNode = parseCode(fileInfo.content)
val filePathWithSlashes = fileInfo.physicalPathAsPath()
val pathWithoutFileEnding = filePathWithSlashes.withoutFileSuffix("py")
val modulePath = pathWithoutFileEnding.parts

val definitions = definitionsQuery.execute(rootNode)
val result = TreeSitterDependencies.analyze(fileInfo.content, Language.PYTHON)
val modulePath = fileInfo.physicalPathAsPath().withoutFileSuffix(PY_SUFFIX).parts

val importsFrom = importQuery.executeImportsFrom(rootNode, fileInfo.content, modulePath)
val wildcardImportsFrom = importQuery.executeWildcardImportsFrom(rootNode, fileInfo.content, modulePath)
val aliasedImportsFrom = importQuery.executeAliasedImportsFrom(rootNode, fileInfo.content, modulePath)
val resolvedImports = result.imports.map { it.resolveRelative(modulePath) }
val regularFromImports = resolvedImports.filter { it.kind == ImportKind.IMPORT_FROM && !it.isAliased && !it.isWildcard }
val wildcardFromImports = resolvedImports.filter { it.kind == ImportKind.IMPORT_FROM && it.isWildcard }
val aliasedFromImports = resolvedImports.filter { it.kind == ImportKind.IMPORT_FROM && it.isAliased }
val standardImports = resolvedImports.filter { it.kind == ImportKind.STANDARD }

val imports = importQuery.executeImports(rootNode, fileInfo.content)
val aliasedImports = importQuery.executeAliasedImports(rootNode, fileInfo.content, modulePath)
val regularFromTwins = regularFromImports.flatMap { it.twinPair() }
val wildcardFromTwins = wildcardFromImports.flatMap { it.twinPair() }
val basePreloadedDeps = regularFromTwins + wildcardFromTwins

val nodes = mutableListOf<Node>()

if (modulePath.last() == "__init__") {
nodes.addAll(
importsFrom.filter { dependency -> "__init__" !in dependency.path.parts }.map { dependency ->
Node(
pathWithName = Path(modulePath + dependency.path.parts.last()),
physicalPath = fileInfo.physicalPath,
language = SupportedLanguage.PYTHON,
nodeType = NodeType.UNKNOWN,
dependencies = importsFrom.toSet(),
usedTypes = setOf(Type.simple(dependency.path.parts.last()))
)
}
)
if (modulePath.lastOrNull() == INIT) {
nodes.addAll(initFileReexportNodes(regularFromImports, regularFromTwins.toSet(), modulePath))
Comment thread
christian-spahn-mw marked this conversation as resolved.
}

if (definitions.isNotEmpty()) {
nodes.addAll(
definitions.map {
extractNodeFromDefinition(
modulePath,
it,
importsFrom + wildcardImportsFrom,
aliasedImportsFrom,
imports,
aliasedImports,
)
}
)
result.declarations.forEach { declaration ->
nodes.add(buildDeclarationNode(declaration, modulePath, basePreloadedDeps, aliasedFromImports, standardImports))
}

return FileReport(nodes)
}

private fun parseCode(pythonCode: String): TSNode {
val parser = TSParser()
parser.language = python
val tree = parser.parseString(null, pythonCode)
return tree.rootNode
}
private fun initFileReexportNodes(
regularFromImports: List<ImportDeclaration>,
regularFromTwins: Set<Dependency>,
modulePath: List<String>
): List<Node> =
regularFromImports.map { import ->
val name = import.path.last()
Node(
pathWithName = Path(modulePath + name),
physicalPath = fileInfo.physicalPath,
language = SupportedLanguage.PYTHON,
nodeType = NodeType.UNKNOWN,
dependencies = regularFromTwins,
usedTypes = setOf(Type.simple(name))
)
}

private fun extractNodeFromDefinition(
private fun buildDeclarationNode(
declaration: Declaration,
modulePath: List<String>,
definition: TSNode,
importFromDependencies: List<Dependency>,
aliasedImportsFrom: Map<String, String>,
imports: List<String>,
aliasedImports: Map<String, String>,
basePreloadedDeps: List<Dependency>,
aliasedFromImports: List<ImportDeclaration>,
standardImports: List<ImportDeclaration>
): Node {
val mutableImportFromDependencies = importFromDependencies.toMutableList()
val nodeBody = nodeAsString(definition, fileInfo.content)
val definitionNode = parseCode(nodeBody).getChild(0)
val pathWithName = Path(modulePath + declaration.name)
val nodeType = declaration.type.toNodeType()

if (definition.type == "identifier") {
if (declaration.type == DeclarationType.VARIABLE) {
return Node(
pathWithName = Path(modulePath + nodeBody),
pathWithName = pathWithName,
physicalPath = fileInfo.physicalPath,
language = SupportedLanguage.PYTHON,
nodeType = NodeType.VARIABLE,
dependencies = setOf(),
usedTypes = setOf()
nodeType = nodeType,
dependencies = emptySet(),
usedTypes = emptySet()
)
}

var classOrFunctionDefinitionNode = definitionNode
if (definition.type == "decorated_definition") {
classOrFunctionDefinitionNode = definitionNode.getChildByFieldName("definition")
}
val (identifierStream, attributeStream) = declaration.usedTypes.partition { it.namespacePrefix.isEmpty() }
val identifierNames = identifierStream.map { it.name }.toSet()

val definitionName = nodeAsString(classOrFunctionDefinitionNode.getChildByFieldName("name"), nodeBody)

val importFromTypes = buildImportFromTypes(
identifierQuery.execute(definitionNode, nodeBody),
aliasedImportsFrom,
mutableImportFromDependencies
)

val importDependencies = buildImportDependencies(imports, aliasedImports, definitionNode, nodeBody)
val deps = basePreloadedDeps.toMutableSet()
aliasedFromImports.forEach { import ->
if (import.path.last() in identifierNames) {
deps.addAll(import.twinPair())
}
}
attributeStream.forEach { ut ->
deps.addAll(matchStandardImportAttribute(ut, standardImports))
}

val usedTypes = identifierStream.map { Type.simple(it.name) }.toSet()
return Node(
pathWithName = Path(modulePath + definitionName),
pathWithName = pathWithName,
physicalPath = fileInfo.physicalPath,
language = SupportedLanguage.PYTHON,
nodeType = nodeType(classOrFunctionDefinitionNode),
dependencies = mutableImportFromDependencies.toSet() + importDependencies,
usedTypes = importFromTypes
nodeType = nodeType,
dependencies = deps,
usedTypes = usedTypes
)
Comment thread
christian-spahn-mw marked this conversation as resolved.
}

private fun nodeType(declaration: TSNode) =
when (declaration.type) {
"class_definition" -> NodeType.CLASS
"function_definition" -> NodeType.FUNCTION
else -> NodeType.UNKNOWN
}

private fun buildImportFromTypes(
types: List<String>,
aliasedImportsFrom: Map<String, String>,
mutableDependencies: MutableList<Dependency>
): Set<Type> =
types
.map {
val aliasedType = checkAliasImportFrom(it, aliasedImportsFrom, mutableDependencies)
Type.simple(aliasedType)
}.toSet()

private fun checkAliasImportFrom(
type: String,
aliasedImportsFrom: Map<String, String>,
mutableDependencies: MutableList<Dependency>
): String {
if (type in aliasedImportsFrom) {
val aliasedImport = aliasedImportsFrom[type]!!.split(".")
val aliasedType = aliasedImport.last()
mutableDependencies.add(Dependency(Path(aliasedImport)))
mutableDependencies.add(Dependency(Path(aliasedImport.dropLast(1) + listOf("__init__", aliasedType))))
return aliasedType
private fun matchStandardImportAttribute(
usedType: UsedType,
standardImports: List<ImportDeclaration>
): List<Dependency> {
val joinedPrefix = usedType.namespacePrefix.joinToString(SEPARATOR)
return standardImports.flatMap { import ->
if (import.path.joinToString(SEPARATOR) == joinedPrefix) {
listOf(
Dependency(Path(import.path + usedType.name)),
Dependency(Path(import.path + INIT + usedType.name))
)
} else {
emptyList()
}
}
}

return type
private fun ImportDeclaration.resolveRelative(modulePath: List<String>): ImportDeclaration {
val first = path.firstOrNull() ?: return this
if (first.isEmpty() || !first.all { it == '.' }) return this
val resolvedPath = modulePath.dropLast(first.length) + path.drop(1)
return copy(path = resolvedPath)
}

private fun buildImportDependencies(
imports: List<String>,
aliasedImports: Map<String, String>,
definitionNode: TSNode,
nodeBody: String
): Set<Dependency> {
val dependencies = mutableSetOf<Dependency>()
if (imports.isNotEmpty() || aliasedImports.isNotEmpty()) {
val attributes = attributeQuery.execute(definitionNode, nodeBody)
attributes.forEach { attribute ->
val attributeList = attribute.split(".").toMutableList()
if (attributeList.size >= 2) {
val attributeType = attributeList.removeLast()
val attributeIdentifier = attributeList.joinToString(".")
if (attributeIdentifier in imports) {
dependencies.add(Dependency(Path(attributeList + listOf(attributeType)), false))
dependencies.add(Dependency(Path(attributeList + listOf("__init__", attributeType)), false))
} else if (attributeIdentifier in aliasedImports) {
val aliasedPath = Path(aliasedImports[attributeIdentifier]!!.split(".") + listOf(attributeType))
dependencies.add(Dependency(aliasedPath, false))
val aliasedInitPath =
Path(aliasedImports[attributeIdentifier]!!.split(".") + listOf("__init__", attributeType))
dependencies.add(Dependency(aliasedInitPath, false))
}
}
}
private fun ImportDeclaration.twinPair(): List<Dependency> {
val canonical = Dependency(Path(path), isWildcard)
val initPath = if (isWildcard) {
path + INIT
} else {
path.dropLast(1) + INIT + path.last()
}
return dependencies
return listOf(canonical, Dependency(Path(initPath), isWildcard))
}
}

This file was deleted.

This file was deleted.

Loading
Loading