Files
Sourcetrail/src/lib/data/Storage.cpp
T
Eberhard Graether cefaeae1ab resource: Updated documentation and smaller fixes
* updated documentation of: cdb loading, advanced settings, graph view
* removed stats of search index
* added global search paths to cdb summary
2016-03-14 14:45:42 +01:00

1097 lines
28 KiB
C++

#include "data/Storage.h"
#include <sstream>
#include <queue>
#include "utility/file/FileSystem.h"
#include "utility/logging/logging.h"
#include "utility/messaging/type/MessageClearErrorCount.h"
#include "utility/messaging/type/MessageShowErrors.h"
#include "utility/TimePoint.h"
#include "utility/utility.h"
#include "utility/utilityString.h"
#include "utility/Version.h"
#include "utility/Cache.h"
#include "data/graph/token_component/TokenComponentAggregation.h"
#include "data/graph/token_component/TokenComponentSignature.h"
#include "data/graph/Graph.h"
#include "data/location/TokenLocation.h"
#include "data/location/TokenLocationFile.h"
#include "data/location/TokenLocationLine.h"
#include "data/parser/ParseLocation.h"
#include "data/type/DataType.h"
#include "settings/ApplicationSettings.h"
Storage::Storage(const FilePath& dbPath)
: m_sqliteStorage(dbPath.str())
{
}
Storage::~Storage()
{
}
Version Storage::getVersion() const
{
return m_sqliteStorage.getVersion();
}
bool Storage::init()
{
m_commandIndex.addNode(NameHierarchy(SearchMatch::getCommandName(SearchMatch::COMMAND_ALL)));
m_commandIndex.addNode(NameHierarchy(SearchMatch::getCommandName(SearchMatch::COMMAND_ERROR)));
return m_sqliteStorage.init();
}
void Storage::clear()
{
m_sqliteStorage.clear();
clearCaches();
}
void Storage::clearCaches()
{
m_tokenIndex.clear();
m_fileNodeIds.clear();
m_hierarchyCache.clear();
}
std::set<FilePath> Storage::getDependingFilePaths(const std::set<FilePath>& filePaths)
{
std::set<FilePath> dependingFilePaths;
for (const FilePath& filePath: filePaths)
{
std::set<FilePath> dependingFilePathsSubset = getDependingFilePaths(filePath);
dependingFilePaths.insert(dependingFilePathsSubset.begin(), dependingFilePathsSubset.end());
}
return dependingFilePaths;
}
std::set<FilePath> Storage::getDependingFilePaths(const FilePath& filePath)
{
std::set<FilePath> dependingFilePaths;
std::vector<StorageEdge> incomingEdges = m_sqliteStorage.getEdgesByTargetType(
getFileNodeId(filePath), Edge::typeToInt(Edge::EDGE_INCLUDE)
);
for (const StorageEdge& incomingEdge: incomingEdges)
{
FilePath dependingFilePath = getFileNodePath(incomingEdge.sourceNodeId);
dependingFilePaths.insert(dependingFilePath);
std::set<FilePath> dependingFilePathsSubset = getDependingFilePaths(dependingFilePath);
dependingFilePaths.insert(dependingFilePathsSubset.begin(), dependingFilePathsSubset.end());
}
return dependingFilePaths;
}
void Storage::clearFileElements(const std::vector<FilePath>& filePaths)
{
std::vector<Id> fileNodeIds;
for (const FilePath& path : filePaths)
{
fileNodeIds.push_back(getFileNodeId(path));
}
if (fileNodeIds.size())
{
m_sqliteStorage.removeElementsWithLocationInFiles(fileNodeIds);
m_sqliteStorage.removeElements(fileNodeIds);
m_sqliteStorage.removeErrorsInFiles(filePaths);
}
}
void Storage::removeUnusedNames() // maybe rename this function. look for callers first.
{
// m_sqliteStorage.removeUnusedNameHierarchyElements();
clearCaches();
}
std::vector<FileInfo> Storage::getInfoOnAllFiles() const
{
std::vector<FileInfo> fileInfos;
std::vector<StorageFile> storageFiles = m_sqliteStorage.getAllFiles();
for (size_t i = 0; i < storageFiles.size(); i++)
{
boost::posix_time::ptime modificationTime = boost::posix_time::not_a_date_time;
if (storageFiles[i].modificationTime != "not-a-date-time")
{
modificationTime = boost::posix_time::time_from_string(storageFiles[i].modificationTime);
}
fileInfos.push_back(FileInfo(
FilePath(storageFiles[i].filePath),
modificationTime
));
}
return fileInfos;
}
const SearchIndex& Storage::getSearchIndex() const
{
return m_tokenIndex;
}
void Storage::logStats() const
{
std::stringstream ss;
StorageStats stats = getStorageStats();
ss << "\nGraph:\n";
ss << "\t" << stats.nodeCount << " Nodes\n";
ss << "\t" << stats.edgeCount << " Edges\n";
ss << "\nSearch:\n";
ss << "\t" << stats.charCount << " Characters\n";
ss << "\t" << stats.wordCount << " Words\n";
ss << "\t" << stats.searchNodeCount << " SearchNodes\n";
ss << "\nCode:\n";
ss << "\t" << stats.fileCount << " Files\n";
ss << "\t" << stats.fileLOCCount << " Lines of Code\n";
ss << "\t" << stats.sourceLocationCount << " Source Locations\n";
LOG_WARNING(ss.str());
}
void Storage::startParsing()
{
MessageClearErrorCount().dispatch();
m_sqliteStorage.setVersion(Version::getApplicationVersion());
}
void Storage::finishParsing()
{
buildSearchIndex();
buildHierarchyCache();
}
void Storage::injectData(std::shared_ptr<IntermediateStorage> injectedStorage)
{
int totalErrorCount = getErrorCount().total;
injectedStorage->transferToStorage(m_sqliteStorage);
if (totalErrorCount != getErrorCount().total)
{
MessageShowErrors msg(getErrorCount());
msg.setSendAsTask(false);
msg.dispatch();
}
}
Id Storage::getIdForNodeWithNameHierarchy(const NameHierarchy& nameHierarchy) const
{
return m_sqliteStorage.getNodeBySerializedName(NameHierarchy::serialize(nameHierarchy)).id;
}
Id Storage::getIdForNodeWithSearchNameHierarchy(const NameHierarchy& nameHierarchy) const
{
SearchNode* node = m_tokenIndex.getNode(nameHierarchy);
if (node)
{
return node->getFirstTokenId();
}
return 0;
}
Id Storage::getIdForEdge(
Edge::EdgeType type, const NameHierarchy& fromNameHierarchy, const NameHierarchy& toNameHierarchy
) const
{
Id sourceId = getIdForNodeWithNameHierarchy(fromNameHierarchy);
Id targetId = getIdForNodeWithNameHierarchy(toNameHierarchy);
return m_sqliteStorage.getEdgeBySourceTargetType(sourceId, targetId, type).id;
}
Id Storage::getIdForFirstNode() const
{
return m_sqliteStorage.getFirstNode().id;
}
NameHierarchy Storage::getNameHierarchyForNodeWithId(Id nodeId) const
{
return NameHierarchy::deserialize(m_sqliteStorage.getNodeById(nodeId).serializedName);
}
Node::NodeType Storage::getNodeTypeForNodeWithId(Id nodeId) const
{
return Node::intToType(m_sqliteStorage.getNodeById(nodeId).type);
}
std::vector<SearchMatch> Storage::getAutocompletionMatches(const std::string& query) const
{
if (query.size() == m_cachedQuery.size() + 1 && query.find(m_cachedQuery) == 0 && m_cachedResults.size())
{
m_cachedResults = m_tokenIndex.runFuzzySearchCached(query, m_cachedResults);
}
else
{
m_cachedResults = m_tokenIndex.runFuzzySearch(query);
}
m_cachedQuery = query;
SearchResults results = m_cachedResults;
SearchResults commandResults = m_commandIndex.runFuzzySearch(query);
results.insert(commandResults.begin(), commandResults.end());
std::vector<SearchMatch> matches = SearchIndex::getMatches(results, query);
LOG_INFO_STREAM(<< matches.size() << " matches for \"" << query << "\"");
if (matches.size() > 100)
{
matches.resize(100);
}
for (SearchMatch& match : matches)
{
if (!match.tokenIds.size())
{
match.searchType = SearchMatch::SEARCH_COMMAND;
match.typeName = "command";
continue;
}
Id elementId = *(match.tokenIds.cbegin());
if (m_sqliteStorage.isNode(elementId))
{
StorageNode node = m_sqliteStorage.getNodeById(elementId);
match.nodeType = Node::intToType(node.type);
match.typeName = Node::getTypeString(match.nodeType);
if (!node.defined && match.nodeType != Node::NODE_UNDEFINED)
{
match.typeName = "undefined " + match.typeName;
}
}
else
{
match.typeName = Edge::getTypeString(Edge::intToType(m_sqliteStorage.getEdgeById(elementId).type));
}
match.searchType = SearchMatch::SEARCH_TOKEN;
}
return matches;
}
std::vector<SearchMatch> Storage::getSearchMatchesForTokenIds(const std::vector<Id>& tokenIds) const
{
std::vector<SearchMatch> matches;
for (Id tokenId : tokenIds)
{
SearchMatch match;
if (m_sqliteStorage.isFile(tokenId))
{
match.nodeType = Node::NODE_FILE;
}
else if (m_sqliteStorage.isNode(tokenId))
{
StorageNode node = m_sqliteStorage.getNodeById(tokenId);
match.nodeType = Node::intToType(node.type);
}
else
{
continue;
}
match.tokenIds.insert(tokenId);
match.nameHierarchy = m_tokenIndex.getNameHierarchyForTokenId(tokenId);
match.searchType = SearchMatch::SEARCH_TOKEN;
matches.push_back(match);
}
return matches;
}
std::shared_ptr<Graph> Storage::getGraphForAll() const
{
std::shared_ptr<Graph> graph = std::make_shared<Graph>();
std::vector<Id> tokenIds;
for (StorageNode node: m_sqliteStorage.getAllNodes())
{
if (node.defined && (!m_hierarchyCache.isChildOfVisibleNodeOrInvisible(node.id) ||
Node::intToType(node.type) == Node::NODE_NAMESPACE))
{
tokenIds.push_back(node.id);
}
}
addNodesToGraph(tokenIds, graph.get());
return graph;
}
std::shared_ptr<Graph> Storage::getGraphForActiveTokenIds(const std::vector<Id>& tokenIds, bool activeOnly) const
{
std::shared_ptr<Graph> g = std::make_shared<Graph>();
Graph* graph = g.get();
if (tokenIds.size() == 1 && !activeOnly)
{
const Id elementId = tokenIds[0];
if (m_sqliteStorage.isNode(elementId))
{
const StorageNode node = m_sqliteStorage.getNodeById(elementId);
addNodeAndAllChildrenToGraph(getLastVisibleParentNodeId(node.id), graph);
std::vector<StorageEdge> edges = m_sqliteStorage.getEdgesBySourceOrTargetId(node.id);
for (size_t i = 0; i < edges.size(); i++)
{
if (Edge::intToType(edges[i].type) != Edge::EDGE_MEMBER)
{
addEdgeAndAllChildrenToGraph(edges[i].id, graph);
}
}
addAggregationEdgesToGraph(elementId, graph);
}
else if (m_sqliteStorage.isEdge(elementId))
{
addEdgeAndAllChildrenToGraph(elementId, graph);
}
}
else if (tokenIds.size() >= 1)
{
for (size_t i = 0; i < tokenIds.size(); i++)
{
const Id elementId = tokenIds[i];
if (m_sqliteStorage.isNode(elementId))
{
addNodeAndAllChildrenToGraph(getLastVisibleParentNodeId(elementId), graph);
}
else
{
addEdgeAndAllChildrenToGraph(elementId, graph);
}
}
}
addComponentAccessToGraph(graph);
return g;
}
std::vector<Id> Storage::getActiveTokenIdsForTokenIds(const std::vector<Id>& tokenIds) const
{
bool different = false;
std::vector<Id> activeIds;
for (Id id : tokenIds)
{
if (m_sqliteStorage.isNode(id))
{
m_hierarchyCache.addFirstVisibleChildIdsForNodeId(id, &activeIds);
if (id != activeIds.back())
{
different = true;
}
}
else
{
activeIds.push_back(id);
}
}
if (!different)
{
return tokenIds;
}
std::set<Id> idSet(activeIds.begin(), activeIds.end());
activeIds.clear();
activeIds.insert(activeIds.end(), idSet.begin(), idSet.end());
return activeIds;
}
// TODO: rename: getActiveElementIdsForId; TODO: make separate function for declarationId
std::vector<Id> Storage::getActiveTokenIdsForId(Id tokenId, Id* declarationId) const
{
std::vector<Id> activeTokenIds;
if (!(m_sqliteStorage.isEdge(tokenId) || m_sqliteStorage.isNode(tokenId)))
{
return activeTokenIds;
}
activeTokenIds.push_back(tokenId);
if (m_sqliteStorage.isNode(tokenId))
{
*declarationId = tokenId;
std::vector<StorageEdge> incomingEdges = m_sqliteStorage.getEdgesByTargetId(tokenId);
for (size_t i = 0; i < incomingEdges.size(); i++)
{
activeTokenIds.push_back(incomingEdges[i].id);
}
}
return activeTokenIds;
}
std::vector<Id> Storage::getNodeIdsForLocationIds(const std::vector<Id>& locationIds) const
{
std::set<Id> nodeIds;
std::set<Id> edgeIds;
for (Id locationId : locationIds)
{
Id elementId = m_sqliteStorage.getElementIdByLocationId(locationId);
StorageEdge edge = m_sqliteStorage.getEdgeById(elementId);
if (edge.id != 0) // here we test if location is an edge.
{
edgeIds.insert(edge.targetNodeId);
}
else
{
nodeIds.insert(elementId);
}
}
if (nodeIds.size())
{
return utility::toVector(nodeIds);
}
return utility::toVector(edgeIds);
}
std::vector<Id> Storage::getTokenIdsForMatches(const std::vector<SearchMatch>& matches) const
{
std::set<Id> idSet;
for (const SearchMatch& match : matches)
{
SearchNode* searchNode = m_tokenIndex.getNode(match.nameHierarchy);
if (searchNode)
{
utility::append(idSet, searchNode->getTokenIds());
}
}
std::vector<Id> ids;
for (std::set<Id>::const_iterator it = idSet.begin(); it != idSet.end(); it++)
{
ids.push_back(*it);
}
return ids;
}
Id Storage::getTokenIdForFileNode(const FilePath& filePath) const
{
return m_sqliteStorage.getFileByPath(filePath.str()).id;
}
std::vector<Id> Storage::getTokenIdsForAggregationEdge(Id sourceId, Id targetId) const
{
std::vector<Id> edgeIds;
std::vector<Id> aggregationEndpointsA = getAllChildNodeIds(sourceId);
std::set<Id> aggregationEndpointsB;
aggregationEndpointsB.insert(targetId);
for (const Id targetChildId: getAllChildNodeIds(targetId))
{
aggregationEndpointsB.insert(targetChildId);
}
for (size_t i = 0; i < aggregationEndpointsA.size(); i++)
{
std::vector<StorageEdge> outgoingEdges = m_sqliteStorage.getEdgesBySourceId(aggregationEndpointsA[i]);
for (size_t j = 0; j < outgoingEdges.size(); j++)
{
if (aggregationEndpointsB.find(outgoingEdges[j].targetNodeId) != aggregationEndpointsB.end())
{
edgeIds.push_back(outgoingEdges[j].id);
}
}
std::vector<StorageEdge> incomingEdges = m_sqliteStorage.getEdgesByTargetId(aggregationEndpointsA[i]);
for (size_t j = 0; j < incomingEdges.size(); j++)
{
if (aggregationEndpointsB.find(incomingEdges[j].sourceNodeId) != aggregationEndpointsB.end())
{
edgeIds.push_back(incomingEdges[j].id);
}
}
}
return edgeIds;
}
std::shared_ptr<TokenLocationCollection> Storage::getTokenLocationsForTokenIds(const std::vector<Id>& tokenIds) const
{
std::shared_ptr<TokenLocationCollection> collection = std::make_shared<TokenLocationCollection>();
std::vector<Id> fileIds;
std::vector<Id> nonFileIds;
for (size_t i = 0; i < tokenIds.size(); i++)
{
if (m_sqliteStorage.isFile(tokenIds[i]))
{
fileIds.push_back(tokenIds[i]);
}
else
{
nonFileIds.push_back(tokenIds[i]);
}
}
for (Id fileId: fileIds)
{
StorageFile storageFile = m_sqliteStorage.getFileById(fileId);
collection->addTokenLocationFileAsPlainCopy(m_sqliteStorage.getTokenLocationsForFile(storageFile.filePath).get());
}
Cache<Id, std::string> filePathCache(
[this](Id id) -> std::string
{
return m_sqliteStorage.getFileById(id).filePath;
}
);
std::vector<StorageSourceLocation> locations = m_sqliteStorage.getTokenLocationsForElementIds(nonFileIds);
for (size_t i = 0; i < locations.size(); i++)
{
const StorageSourceLocation& location = locations[i];
std::string filePath = filePathCache.getValue(location.fileNodeId);
TokenLocation* loc = collection->addTokenLocation(
location.id,
location.elementId,
filePath,
location.startLine,
location.startCol,
location.endLine,
location.endCol
);
if (loc)
{
loc->setType(location.isScope ? TokenLocation::LOCATION_SCOPE : TokenLocation::LOCATION_TOKEN);
}
}
return collection;
}
std::shared_ptr<TokenLocationCollection> Storage::getTokenLocationsForLocationIds(const std::vector<Id>& locationIds) const
{
std::shared_ptr<TokenLocationCollection> collection = std::make_shared<TokenLocationCollection>();
for (size_t i = 0; i < locationIds.size(); i++)
{
StorageSourceLocation location = m_sqliteStorage.getSourceLocationById(locationIds[i]);
collection->addTokenLocation(
location.id,
location.elementId,
m_sqliteStorage.getFileById(location.fileNodeId).filePath, // TODO: optimize: only once per file!
location.startLine,
location.startCol,
location.endLine,
location.endCol)->setType(location.isScope ? TokenLocation::LOCATION_SCOPE : TokenLocation::LOCATION_TOKEN
);
}
return collection;
}
std::shared_ptr<TokenLocationFile> Storage::getTokenLocationsForFile(const std::string& filePath) const
{
std::shared_ptr<TokenLocationFile> locationFile = m_sqliteStorage.getTokenLocationsForFile(filePath);
locationFile->isWholeCopy = true;
return locationFile;
}
std::shared_ptr<TokenLocationFile> Storage::getTokenLocationsForLinesInFile(
const std::string& filePath, uint firstLineNumber, uint lastLineNumber
) const
{
return m_sqliteStorage.getTokenLocationsForFile(filePath)->getFilteredByLines(firstLineNumber, lastLineNumber);
}
TokenLocationCollection Storage::getErrorTokenLocations(std::vector<std::string>* errorMessages) const
{
TokenLocationCollection errorCollection;
std::vector<StorageError> errors = m_sqliteStorage.getAllErrors();
for (size_t i = 0; i < errors.size(); i++)
{
const StorageError& error = errors[i];
errorCollection.addTokenLocation(
i, i, error.filePath, error.lineNumber, error.columnNumber, error.lineNumber, error.columnNumber);
errorMessages->push_back(error.message);
}
return errorCollection;
}
std::shared_ptr<TokenLocationFile> Storage::getCommentLocationsInFile(const FilePath& filePath) const
{
std::shared_ptr<TokenLocationFile> file = std::make_shared<TokenLocationFile>(filePath);
std::vector<StorageCommentLocation> storageLocations = m_sqliteStorage.getCommentLocationsInFile(filePath);
for (size_t i = 0; i < storageLocations.size(); i++)
{
file->addTokenLocation(
storageLocations[i].id,
0, // comment token location has no element.
storageLocations[i].startLine,
storageLocations[i].startCol,
storageLocations[i].endLine,
storageLocations[i].endCol
);
}
return file;
}
std::shared_ptr<TextAccess> Storage::getFileContent(const FilePath& filePath) const
{
return m_sqliteStorage.getFileContentByPath(filePath.str());
}
TimePoint Storage::getFileModificationTime(const FilePath& filePath) const
{
return TimePoint(m_sqliteStorage.getFileByPath(filePath.str()).modificationTime);
}
ErrorCountInfo Storage::getErrorCount() const
{
return ErrorCountInfo(m_sqliteStorage.getAllErrors().size(), m_sqliteStorage.getFatalErrors().size());
}
StorageStats Storage::getStorageStats() const
{
StorageStats stats;
stats.nodeCount = m_sqliteStorage.getNodeCount();
stats.edgeCount = m_sqliteStorage.getEdgeCount();
// Takes too much time
// stats.charCount = m_tokenIndex.getCharCount();
// stats.wordCount = m_tokenIndex.getWordCount();
// stats.searchNodeCount = m_tokenIndex.getNodeCount();
stats.fileCount = m_sqliteStorage.getFileCount();
stats.fileLOCCount = m_sqliteStorage.getFileLOCCount();
stats.sourceLocationCount = m_sqliteStorage.getSourceLocationCount();
stats.errorCount = getErrorCount();
return stats;
}
Id Storage::addNodeHierarchy(Node::NodeType nodeType, NameHierarchy nameHierarchy, bool defined)
{
if (nameHierarchy.size() == 0)
{
return 0;
}
Id parentNodeId = 0;
bool nodeMayExist = true;
NameHierarchy currentNameHierarchy;
for (size_t i = 0; i < nameHierarchy.size(); i++)
{
currentNameHierarchy.push(nameHierarchy[i]);
const bool isLastElement = (i == nameHierarchy.size() - 1);
Node::NodeType type = (isLastElement ? nodeType : Node::NODE_UNDEFINED);
StorageNode node(0, 0, "", false);
if (nodeMayExist)
{
node = m_sqliteStorage.getNodeBySerializedName(NameHierarchy::serialize(currentNameHierarchy));
}
Id nodeId = node.id;
if (nodeId && !node.defined && isLastElement && defined) // todo: move this down!
{
m_sqliteStorage.setNodeDefined(true, nodeId);
}
if (nodeId == 0)
{
nodeMayExist = false;
nodeId = m_sqliteStorage.addNode(Node::typeToInt(type), NameHierarchy::serialize(currentNameHierarchy), isLastElement && defined);
if (parentNodeId != 0)
{
addEdge(parentNodeId, nodeId, Edge::EDGE_MEMBER);
}
}
else if (isLastElement) // Update the type of the last node if the new type is more specific.
{
Node::NodeType storedType = Node::intToType(node.type);
if (!node.defined && type > storedType)
{
m_sqliteStorage.setNodeType(Node::typeToInt(type), nodeId);
}
}
parentNodeId = nodeId;
}
return parentNodeId;
}
Id Storage::addSourceLocation(Id elementNodeId, const ParseLocation &location, bool isScope)
{
if (!location.isValid())
{
return 0;
}
if (location.filePath.empty())
{
LOG_ERROR("no filename set!");
return 0;
}
else
{
Id fileNodeId = getFileNodeId(location.filePath);
if (!fileNodeId)
{
LOG_ERROR("Can't create source location, file node does not exist for: " + location.filePath.str());
return 0;
}
Id locationId = m_sqliteStorage.addSourceLocation(
elementNodeId, fileNodeId, location.startLineNumber, location.startColumnNumber,
location.endLineNumber, location.endColumnNumber, isScope
);
return locationId;
}
}
Id Storage::addEdge(Id sourceNodeId, Id targetNodeId, Edge::EdgeType type)
{
if (!sourceNodeId || !targetNodeId)
{
return 0;
}
Id edgeId = m_sqliteStorage.getEdgeBySourceTargetType(sourceNodeId, targetNodeId, type).id;
if (!edgeId)
{
edgeId = m_sqliteStorage.addEdge(type, sourceNodeId, targetNodeId);
}
return edgeId;
}
Id Storage::addEdge(Id sourceNodeId, Id targetNodeId, Edge::EdgeType type, ParseLocation location)
{
if (!sourceNodeId || !targetNodeId)
{
return 0;
}
Id edgeId = addEdge(sourceNodeId, targetNodeId, type);
addSourceLocation(edgeId, location, false);
return edgeId;
}
Id Storage::getFileNodeId(const FilePath& filePath) const
{
std::map<FilePath, Id>::const_iterator it = m_fileNodeIds.find(filePath);
if (it != m_fileNodeIds.end())
{
return it->second;
}
if (filePath.empty())
{
LOG_ERROR("No file path set");
return 0;
}
StorageFile storageFile = m_sqliteStorage.getFileByPath(filePath.str());
if (storageFile.id == 0)
{
return 0;
}
m_fileNodeIds.emplace(filePath, storageFile.id);
return storageFile.id;
}
FilePath Storage::getFileNodePath(Id fileId) const
{
for (const std::pair<FilePath, Id>& p : m_fileNodeIds)
{
if (p.second == fileId)
{
return p.first;
}
}
return m_sqliteStorage.getFileById(fileId).filePath;
}
Id Storage::getLastVisibleParentNodeId(const Id nodeId) const
{
return m_hierarchyCache.getLastVisibleParentNodeId(nodeId);
}
std::vector<Id> Storage::getAllChildNodeIds(const Id nodeId) const
{
std::vector<Id> childNodeIds;
std::vector<Id> edgeIds;
m_hierarchyCache.addAllChildIdsForNodeId(nodeId, &childNodeIds, &edgeIds);
return childNodeIds;
}
void Storage::addEdgeAndAllChildrenToGraph(const Id edgeId, Graph* graph) const
{
StorageEdge storageEdge = m_sqliteStorage.getEdgeById(edgeId);
Node* sourceNode = graph->getNodeById(storageEdge.sourceNodeId);
Node* targetNode = graph->getNodeById(storageEdge.targetNodeId);
if (!sourceNode)
{
addNodeAndAllChildrenToGraph(getLastVisibleParentNodeId(storageEdge.sourceNodeId), graph);
sourceNode = graph->getNodeById(storageEdge.sourceNodeId);
}
if (!targetNode)
{
addNodeAndAllChildrenToGraph(getLastVisibleParentNodeId(storageEdge.targetNodeId), graph);
targetNode = graph->getNodeById(storageEdge.targetNodeId);
}
graph->createEdge(edgeId, Edge::intToType(storageEdge.type), sourceNode, targetNode);
}
Node* Storage::addNodeAndAllChildrenToGraph(const Id nodeId, Graph* graph) const
{
Node* node = graph->getNodeById(nodeId);
if (node)
{
return node;
}
std::vector<Id> nodeIdsToAdd;
std::vector<Id> edgeIdsToAdd;
nodeIdsToAdd.push_back(nodeId);
m_hierarchyCache.addAllChildIdsForNodeId(nodeId, &nodeIdsToAdd, &edgeIdsToAdd);
addNodesToGraph(nodeIdsToAdd, graph);
addEdgesToGraph(edgeIdsToAdd, graph);
return graph->getNodeById(nodeId);
}
void Storage::addAggregationEdgesToGraph(const Id nodeId, Graph* graph) const
{
struct EdgeInfo
{
Id edgeId;
bool forward;
};
// build aggregation edges:
// get all children of the active node
std::vector<Id> childNodeIds = getAllChildNodeIds(nodeId);
// get all edges of the children
std::map<Id, std::vector<EdgeInfo>> connectedNodeIds;
std::vector<StorageEdge> outgoingEdges = m_sqliteStorage.getEdgesBySourceIds(childNodeIds);
for (size_t j = 0; j < outgoingEdges.size(); j++)
{
EdgeInfo edgeInfo;
edgeInfo.edgeId = outgoingEdges[j].id;
edgeInfo.forward = true;
connectedNodeIds[outgoingEdges[j].targetNodeId].push_back(edgeInfo);
}
std::vector<StorageEdge> incomingEdges = m_sqliteStorage.getEdgesByTargetIds(childNodeIds);
for (size_t j = 0; j < incomingEdges.size(); j++)
{
EdgeInfo edgeInfo;
edgeInfo.edgeId = incomingEdges[j].id;
edgeInfo.forward = false;
connectedNodeIds[incomingEdges[j].sourceNodeId].push_back(edgeInfo);
}
// get all parent nodes of all connected nodes (up to last level except namespace/undefined)
Id nodeParentNodeId = getLastVisibleParentNodeId(nodeId);
std::map<Id, std::vector<EdgeInfo>> connectedParentNodeIds;
for (const std::pair<Id, std::vector<EdgeInfo>>& p : connectedNodeIds)
{
Id parentNodeId = getLastVisibleParentNodeId(p.first);
if (parentNodeId != nodeParentNodeId)
{
utility::append(connectedParentNodeIds[parentNodeId], p.second);
}
}
// add hierarchies for these parents and create aggregation edges between parents and active node
Node* sourceNode = graph->getNodeById(nodeId);
for (const std::pair<Id, std::vector<EdgeInfo>> p : connectedParentNodeIds)
{
const Id aggregationTargetNodeId = p.first;
Node* targetNode = graph->getNodeById(aggregationTargetNodeId);
if (!targetNode)
{
targetNode = addNodeAndAllChildrenToGraph(aggregationTargetNodeId, graph);
}
std::shared_ptr<TokenComponentAggregation> componentAggregation = std::make_shared<TokenComponentAggregation>();
for (const EdgeInfo& edgeInfo: p.second)
{
componentAggregation->addAggregationId(edgeInfo.edgeId, edgeInfo.forward);
}
Edge* edge = graph->createEdge(
*componentAggregation->getAggregationIds().begin(),
Edge::EDGE_AGGREGATION,
sourceNode,
targetNode
);
edge->addComponentAggregation(componentAggregation);
}
}
void Storage::addNodesToGraph(const std::vector<Id> nodeIds, Graph* graph) const
{
std::vector<StorageNode> storageNodes = m_sqliteStorage.getNodesByIds(nodeIds);
for (const StorageNode& storageNode : storageNodes)
{
NameHierarchy nameHierarchy = NameHierarchy::deserialize(storageNode.serializedName);
Node::NodeType type = Node::intToType(storageNode.type);
Node* node = graph->createNode(
storageNode.id,
type,
nameHierarchy,
storageNode.defined
);
if (type == Node::NODE_FUNCTION || type == Node::NODE_METHOD)
{
std::string signatureString = nameHierarchy.getRawNameWithSignature();
if (signatureString.size() > 0) // this should always be the case since functions and methods must have sigs.
{
node->addComponentSignature(
std::make_shared<TokenComponentSignature>(signatureString)
);
}
}
}
}
void Storage::addEdgesToGraph(const std::vector<Id> edgeIds, Graph* graph) const
{
std::vector<StorageEdge> storageEdges = m_sqliteStorage.getEdgesByIds(edgeIds);
for (const StorageEdge& storageEdge : storageEdges)
{
Node* sourceNode = graph->getNodeById(storageEdge.sourceNodeId);
Node* targetNode = graph->getNodeById(storageEdge.targetNodeId);
if (sourceNode && targetNode)
{
graph->createEdge(storageEdge.id, Edge::intToType(storageEdge.type), sourceNode, targetNode);
}
else
{
LOG_ERROR("Can't add edge because nodes are not present");
}
}
}
void Storage::addComponentAccessToGraph(Graph* graph) const
{
std::vector<Id> memberEdgeIds;
graph->forEachEdge(
[&memberEdgeIds](Edge* edge)
{
if (!edge->isType(Edge::EDGE_MEMBER))
{
return;
}
memberEdgeIds.push_back(edge->getId());
}
);
std::vector<StorageComponentAccess> accesses = m_sqliteStorage.getComponentAccessByMemberEdgeIds(memberEdgeIds);
for (const StorageComponentAccess& access : accesses)
{
if (access.memberEdgeId && access.type)
{
graph->getEdgeById(access.memberEdgeId)->addComponentAccess(
std::make_shared<TokenComponentAccess>(TokenComponentAccess::intToType(access.type)));
}
}
}
void Storage::buildSearchIndex()
{
for (StorageNode node: m_sqliteStorage.getAllNodes())
{
m_tokenIndex.addTokenId(m_tokenIndex.addNode(NameHierarchy::deserialize(node.serializedName)), node.id);
}
}
void Storage::buildHierarchyCache()
{
std::vector<StorageEdge> memberEdges = m_sqliteStorage.getEdgesByType(Edge::typeToInt(Edge::EDGE_MEMBER));
Cache<Id, Node::NodeType> nodeTypeCache([this](Id id){
return Node::intToType(m_sqliteStorage.getNodeById(id).type);
});
for (const StorageEdge& edge : memberEdges)
{
bool isVisible = !(nodeTypeCache.getValue(edge.sourceNodeId) & Node::NODE_NOT_VISIBLE);
m_hierarchyCache.createConnection(edge.id, edge.sourceNodeId, edge.targetNodeId, isVisible);
}
}