Matcher*
DefDelWriter_Seg_Deletions_IMP(DefaultDeletionsWriter *self,
SegReader *seg_reader) {
DefaultDeletionsWriterIVARS *const ivars = DefDelWriter_IVARS(self);
Matcher *deletions = NULL;
Segment *segment = SegReader_Get_Segment(seg_reader);
String *seg_name = Seg_Get_Name(segment);
Integer *tick_obj = (Integer*)Hash_Fetch(ivars->name_to_tick,
seg_name);
size_t tick = tick_obj ? (size_t)Int_Get_Value(tick_obj) : 0;
SegReader *candidate = tick_obj
? (SegReader*)Vec_Fetch(ivars->seg_readers, tick)
: NULL;
if (tick_obj) {
DeletionsReader *del_reader
= (DeletionsReader*)SegReader_Obtain(
candidate, Class_Get_Name(DELETIONSREADER));
if (ivars->updated[tick] || DelReader_Del_Count(del_reader)) {
BitVector *deldocs = (BitVector*)Vec_Fetch(ivars->bit_vecs, tick);
deletions = (Matcher*)BitVecMatcher_new(deldocs);
}
}
else { // Sanity check.
THROW(ERR, "Couldn't find SegReader %o", seg_reader);
}
return deletions;
}
void
Freezer_serialize(Obj *obj, OutStream *outstream) {
if (Obj_is_a(obj, STRING)) {
Freezer_serialize_string((String*)obj, outstream);
}
else if (Obj_is_a(obj, BLOB)) {
Freezer_serialize_blob((Blob*)obj, outstream);
}
else if (Obj_is_a(obj, VECTOR)) {
Freezer_serialize_varray((Vector*)obj, outstream);
}
else if (Obj_is_a(obj, HASH)) {
Freezer_serialize_hash((Hash*)obj, outstream);
}
else if (Obj_is_a(obj, INTEGER)) {
int64_t val = Int_Get_Value((Integer*)obj);
OutStream_Write_C64(outstream, (uint64_t)val);
}
else if (Obj_is_a(obj, FLOAT)) {
double val = Float_Get_Value((Float*)obj);
OutStream_Write_F64(outstream, val);
}
else if (Obj_is_a(obj, BOOLEAN)) {
bool val = Bool_Get_Value((Boolean*)obj);
OutStream_Write_U8(outstream, (uint8_t)val);
}
else if (Obj_is_a(obj, QUERY)) {
Query_Serialize((Query*)obj, outstream);
}
else if (Obj_is_a(obj, DOC)) {
Doc_Serialize((Doc*)obj, outstream);
}
else if (Obj_is_a(obj, DOCVECTOR)) {
DocVec_Serialize((DocVector*)obj, outstream);
}
else if (Obj_is_a(obj, TERMVECTOR)) {
TV_Serialize((TermVector*)obj, outstream);
}
else if (Obj_is_a(obj, SIMILARITY)) {
Sim_Serialize((Similarity*)obj, outstream);
}
else if (Obj_is_a(obj, MATCHDOC)) {
MatchDoc_Serialize((MatchDoc*)obj, outstream);
}
else if (Obj_is_a(obj, TOPDOCS)) {
TopDocs_Serialize((TopDocs*)obj, outstream);
}
else if (Obj_is_a(obj, SORTSPEC)) {
SortSpec_Serialize((SortSpec*)obj, outstream);
}
else if (Obj_is_a(obj, SORTRULE)) {
SortRule_Serialize((SortRule*)obj, outstream);
}
else {
THROW(ERR, "Don't know how to serialize a %o",
Obj_get_class_name(obj));
}
}
int32_t
DefDelWriter_Seg_Del_Count_IMP(DefaultDeletionsWriter *self,
String *seg_name) {
DefaultDeletionsWriterIVARS *const ivars = DefDelWriter_IVARS(self);
Integer *tick = (Integer*)Hash_Fetch(ivars->name_to_tick, seg_name);
BitVector *deldocs = tick
? (BitVector*)Vec_Fetch(ivars->bit_vecs, (size_t)Int_Get_Value(tick))
: NULL;
return deldocs ? (int32_t)BitVec_Count(deldocs) : 0;
}
int32_t
Seg_Field_Num_IMP(Segment *self, String *field) {
if (field == NULL) {
return 0;
}
else {
SegmentIVARS *const ivars = Seg_IVARS(self);
Integer *num = (Integer*)Hash_Fetch(ivars->by_name, field);
return num ? (int32_t)Int_Get_Value(num) : 0;
}
}
int32_t
Seg_Add_Field_IMP(Segment *self, String *field) {
SegmentIVARS *const ivars = Seg_IVARS(self);
Integer *num = (Integer*)Hash_Fetch(ivars->by_name, field);
if (num) {
return (int32_t)Int_Get_Value(num);
}
else {
int32_t field_num = (int32_t)Vec_Get_Size(ivars->by_num);
Hash_Store(ivars->by_name, field, (Obj*)Int_new(field_num));
Vec_Push(ivars->by_num, (Obj*)Str_Clone(field));
return field_num;
}
}
void
DocWriter_Add_Inverted_Doc_IMP(DocWriter *self, Inverter *inverter,
int32_t doc_id) {
DocWriterIVARS *const ivars = DocWriter_IVARS(self);
OutStream *dat_out = S_lazy_init(self);
OutStream *ix_out = ivars->ix_out;
uint32_t num_stored = 0;
int64_t start = OutStream_Tell(dat_out);
int64_t expected = OutStream_Tell(ix_out) / 8;
// Verify doc id.
if (doc_id != expected) {
THROW(ERR, "Expected doc id %i64 but got %i32", expected, doc_id);
}
// Write the number of stored fields.
Inverter_Iterate(inverter);
while (Inverter_Next(inverter)) {
FieldType *type = Inverter_Get_Type(inverter);
if (FType_Stored(type)) { num_stored++; }
}
OutStream_Write_CU32(dat_out, num_stored);
Inverter_Iterate(inverter);
while (Inverter_Next(inverter)) {
// Only store fields marked as "stored".
FieldType *type = Inverter_Get_Type(inverter);
if (FType_Stored(type)) {
String *field = Inverter_Get_Field_Name(inverter);
Obj *value = Inverter_Get_Value(inverter);
Freezer_serialize_string(field, dat_out);
switch (FType_Primitive_ID(type) & FType_PRIMITIVE_ID_MASK) {
case FType_TEXT: {
const char *buf = Str_Get_Ptr8((String*)value);
size_t size = Str_Get_Size((String*)value);
if (size > INT32_MAX) {
THROW(ERR, "Field %o over 2GB: %u64", field,
(uint64_t)size);
}
OutStream_Write_CU32(dat_out, (uint32_t)size);
OutStream_Write_Bytes(dat_out, buf, size);
break;
}
case FType_BLOB: {
const char *buf = Blob_Get_Buf((Blob*)value);
size_t size = Blob_Get_Size((Blob*)value);
if (size > INT32_MAX) {
THROW(ERR, "Field %o over 2GB: %u64", field,
(uint64_t)size);
}
OutStream_Write_CU32(dat_out, (uint32_t)size);
OutStream_Write_Bytes(dat_out, buf, size);
break;
}
case FType_INT32: {
int32_t val = (int32_t)Int_Get_Value((Integer*)value);
OutStream_Write_CI32(dat_out, val);
break;
}
case FType_INT64: {
int64_t val = Int_Get_Value((Integer*)value);
OutStream_Write_CI64(dat_out, val);
break;
}
case FType_FLOAT32: {
float val = (float)Float_Get_Value((Float*)value);
OutStream_Write_F32(dat_out, val);
break;
}
case FType_FLOAT64: {
double val = Float_Get_Value((Float*)value);
OutStream_Write_F64(dat_out, val);
break;
}
default:
THROW(ERR, "Unrecognized type: %o", type);
}
}
}
// Write file pointer.
OutStream_Write_I64(ix_out, start);
}
static void
S_write_val(Obj *val, int8_t prim_id, OutStream *ix_out, OutStream *dat_out,
int64_t dat_start) {
if (val) {
switch (prim_id & FType_PRIMITIVE_ID_MASK) {
case FType_TEXT: {
String *string = (String*)val;
int64_t dat_pos = OutStream_Tell(dat_out) - dat_start;
OutStream_Write_I64(ix_out, dat_pos);
OutStream_Write_Bytes(dat_out, Str_Get_Ptr8(string),
Str_Get_Size(string));
break;
}
case FType_BLOB: {
Blob *blob = (Blob*)val;
int64_t dat_pos = OutStream_Tell(dat_out) - dat_start;
OutStream_Write_I64(ix_out, dat_pos);
OutStream_Write_Bytes(dat_out, Blob_Get_Buf(blob),
Blob_Get_Size(blob));
break;
}
case FType_INT32: {
int32_t i32 = (int32_t)Int_Get_Value((Integer*)val);
OutStream_Write_I32(dat_out, i32);
break;
}
case FType_INT64: {
int64_t i64 = Int_Get_Value((Integer*)val);
OutStream_Write_I64(dat_out, i64);
break;
}
case FType_FLOAT32: {
float f32 = (float)Float_Get_Value((Float*)val);
OutStream_Write_F32(dat_out, f32);
break;
}
case FType_FLOAT64: {
double f64 = Float_Get_Value((Float*)val);
OutStream_Write_F64(dat_out, f64);
break;
}
default:
THROW(ERR, "Unrecognized primitive id: %i32", (int32_t)prim_id);
}
}
else {
switch (prim_id & FType_PRIMITIVE_ID_MASK) {
case FType_TEXT:
case FType_BLOB: {
int64_t dat_pos = OutStream_Tell(dat_out) - dat_start;
OutStream_Write_I64(ix_out, dat_pos);
}
break;
case FType_INT32:
OutStream_Write_I32(dat_out, 0);
break;
case FType_INT64:
OutStream_Write_I64(dat_out, 0);
break;
case FType_FLOAT64:
OutStream_Write_F64(dat_out, 0.0);
break;
case FType_FLOAT32:
OutStream_Write_F32(dat_out, 0.0f);
break;
default:
THROW(ERR, "Unrecognized primitive id: %i32", (int32_t)prim_id);
}
}
}
