end/officeconvert
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

allow specifying conversion resolution, drop explicit dpi

Browse Source
This commit is contained in:
Elijah Duffy
2026-03-27 13:51:56 -07:00
parent 5f68aa5567
commit 5923bff155
14 changed files with 398 additions and 94 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.env.example
-1
View File
@@ -4,7 +4,6 @@ S3_USE_SSL=false
S3_ACCESS_KEY=minioadmin
S3_SECRET_KEY=minioadmin
S3_SESSION_TTL_SECONDS=3600
CONVERSION_IMAGE_DPI=72
CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS=180
CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS=1800
CONVERSION_PPTX_TO_PDF_BASE_TIMEOUT_SECONDS=45
Makefile
-1
View File
@@ -35,7 +35,6 @@ run-server:
export S3_ACCESS_KEY="$${S3_ACCESS_KEY:-minioadmin}"; \
export S3_SECRET_KEY="$${S3_SECRET_KEY:-minioadmin}"; \
export S3_SESSION_TTL_SECONDS="$${S3_SESSION_TTL_SECONDS:-3600}"; \
export CONVERSION_IMAGE_DPI="$${CONVERSION_IMAGE_DPI:-72}"; \
export CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS="$${CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS:-180}"; \
export CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS="$${CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS:-1800}"; \
export CONVERSION_PPTX_TO_PDF_BASE_TIMEOUT_SECONDS="$${CONVERSION_PPTX_TO_PDF_BASE_TIMEOUT_SECONDS:-45}"; \
README.md
+5 -3
View File
@@ -94,7 +94,7 @@ make run-server
- defaults S3 endpoint to `localhost:8333` for host-based development
- auto-normalizes `seaweedfs:8333` to `localhost:8333` for host runs
- supports optional `UVICORN_HOST` and `UVICORN_PORT` overrides
- exposes conversion tuning vars (`CONVERSION_IMAGE_DPI`, `CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS`, `CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS`)
- exposes conversion timeout tuning vars (`CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS`, `CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS`)
Server endpoint base URL:
@@ -107,7 +107,7 @@ Create a conversion request:
```bash
curl \
--header "Content-Type: application/json" \
--data '{"sourceFilename":"example.pptx"}' \
--data '{"sourceFilename":"example.pptx","resolution":"CONVERSION_RESOLUTION_FHD"}' \
http://localhost:8080/officeconvertapi.v1.ConversionService/CreateConversion
```
@@ -139,6 +139,8 @@ Use `.env.example` as your baseline env configuration.
If conversion fails on larger decks, tune these environment variables:
- `CONVERSION_IMAGE_DPI` (default `72`): lower values reduce image generation time.
- `CreateConversionRequest.resolution` controls output dimensions via presets: `SD`, `HD`, `FHD`, `QHD`, `UHD`.
- Omitting `resolution` (or sending `CONVERSION_RESOLUTION_UNSPECIFIED`) defaults to `FHD`.
- Rasterization DPI is inferred automatically from source slide size and selected output dimensions.
- `CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS` (default `180`): timeout for LibreOffice export.
- `CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS` (default `1800`): timeout for Poppler rasterization.
clients/go/officeconvertclient/client.go
+2
View File
@@ -46,9 +46,11 @@ func (c *Client) SetPollInterval(interval time.Duration) {
func (c *Client) CreateConversion(
ctx context.Context,
sourceFilename string,
resolution officeconvertapiv1.ConversionResolution,
) (*officeconvertapiv1.CreateConversionResponse, error) {
req := connect.NewRequest(&officeconvertapiv1.CreateConversionRequest{
SourceFilename: sourceFilename,
Resolution: resolution,
})
res, err := c.rpc.CreateConversion(ctx, req)
if err != nil {
clients/go/officeconvertclient/orchestrate.go
+5 -1
View File
@@ -16,7 +16,11 @@ type ConversionResult struct {
// ConvertPPTXFile runs the full create-upload-start-wait-fetch flow.
func (c *Client) ConvertPPTXFile(ctx context.Context, localPPTXPath string) (*ConversionResult, error) {
createRes, err := c.CreateConversion(ctx, filepath.Base(localPPTXPath))
createRes, err := c.CreateConversion(
ctx,
filepath.Base(localPPTXPath),
officeconvertapiv1.ConversionResolution_CONVERSION_RESOLUTION_UNSPECIFIED,
)
if err != nil {
return nil, fmt.Errorf("create conversion: %w", err)
}
deploy/docker-compose.yml
-1
View File
@@ -27,7 +27,6 @@ services:
S3_ACCESS_KEY: ${S3_ACCESS_KEY:-minioadmin}
S3_SECRET_KEY: ${S3_SECRET_KEY:-minioadmin}
S3_SESSION_TTL_SECONDS: ${S3_SESSION_TTL_SECONDS:-3600}
CONVERSION_IMAGE_DPI: ${CONVERSION_IMAGE_DPI:-72}
CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS: ${CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS:-180}
CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS: ${CONVERSION_PDF_TO_IMAGES_TIMEOUT_SECONDS:-1800}
CONVERSION_PPTX_TO_PDF_BASE_TIMEOUT_SECONDS: ${CONVERSION_PPTX_TO_PDF_BASE_TIMEOUT_SECONDS:-45}
gen/go/officeconvertapi/v1/conversion.pb.go
+139 -42
View File
@@ -137,6 +137,65 @@ func (ConversionPhase) EnumDescriptor() ([]byte, []int) {
return file_officeconvertapi_v1_conversion_proto_rawDescGZIP(), []int{1}
}
// ConversionResolution represents preset output quality targets.
type ConversionResolution int32
const (
ConversionResolution_CONVERSION_RESOLUTION_UNSPECIFIED ConversionResolution = 0
ConversionResolution_CONVERSION_RESOLUTION_SD ConversionResolution = 1
ConversionResolution_CONVERSION_RESOLUTION_HD ConversionResolution = 2
ConversionResolution_CONVERSION_RESOLUTION_FHD ConversionResolution = 3
ConversionResolution_CONVERSION_RESOLUTION_QHD ConversionResolution = 4
ConversionResolution_CONVERSION_RESOLUTION_UHD ConversionResolution = 5
)
// Enum value maps for ConversionResolution.
var (
ConversionResolution_name = map[int32]string{
0: "CONVERSION_RESOLUTION_UNSPECIFIED",
1: "CONVERSION_RESOLUTION_SD",
2: "CONVERSION_RESOLUTION_HD",
3: "CONVERSION_RESOLUTION_FHD",
4: "CONVERSION_RESOLUTION_QHD",
5: "CONVERSION_RESOLUTION_UHD",
}
ConversionResolution_value = map[string]int32{
"CONVERSION_RESOLUTION_UNSPECIFIED": 0,
"CONVERSION_RESOLUTION_SD": 1,
"CONVERSION_RESOLUTION_HD": 2,
"CONVERSION_RESOLUTION_FHD": 3,
"CONVERSION_RESOLUTION_QHD": 4,
"CONVERSION_RESOLUTION_UHD": 5,
}
)
func (x ConversionResolution) Enum() *ConversionResolution {
p := new(ConversionResolution)
*p = x
return p
}
func (x ConversionResolution) String() string {
return protoimpl.X.EnumStringOf(x.Descriptor(), protoreflect.EnumNumber(x))
}
func (ConversionResolution) Descriptor() protoreflect.EnumDescriptor {
return file_officeconvertapi_v1_conversion_proto_enumTypes[2].Descriptor()
}
func (ConversionResolution) Type() protoreflect.EnumType {
return &file_officeconvertapi_v1_conversion_proto_enumTypes[2]
}
func (x ConversionResolution) Number() protoreflect.EnumNumber {
return protoreflect.EnumNumber(x)
}
// Deprecated: Use ConversionResolution.Descriptor instead.
func (ConversionResolution) EnumDescriptor() ([]byte, []int) {
return file_officeconvertapi_v1_conversion_proto_rawDescGZIP(), []int{2}
}
// Slide contains extracted notes and the rendered image URL for one slide.
type Slide struct {
state protoimpl.MessageState `protogen:"open.v1"`
@@ -205,6 +264,8 @@ type SlideDeck struct {
SourceFilename string `protobuf:"bytes,2,opt,name=source_filename,json=sourceFilename,proto3" json:"source_filename,omitempty"`
Slides []*Slide `protobuf:"bytes,3,rep,name=slides,proto3" json:"slides,omitempty"`
CreatedAt *timestamppb.Timestamp `protobuf:"bytes,4,opt,name=created_at,json=createdAt,proto3" json:"created_at,omitempty"`
Width int32 `protobuf:"varint,5,opt,name=width,proto3" json:"width,omitempty"`
Height int32 `protobuf:"varint,6,opt,name=height,proto3" json:"height,omitempty"`
unknownFields protoimpl.UnknownFields
sizeCache protoimpl.SizeCache
}
@@ -267,10 +328,25 @@ func (x *SlideDeck) GetCreatedAt() *timestamppb.Timestamp {
return nil
}
func (x *SlideDeck) GetWidth() int32 {
if x != nil {
return x.Width
}
return 0
}
func (x *SlideDeck) GetHeight() int32 {
if x != nil {
return x.Height
}
return 0
}
// CreateConversionRequest starts a conversion session.
type CreateConversionRequest struct {
state protoimpl.MessageState `protogen:"open.v1"`
SourceFilename string `protobuf:"bytes,1,opt,name=source_filename,json=sourceFilename,proto3" json:"source_filename,omitempty"`
Resolution ConversionResolution `protobuf:"varint,2,opt,name=resolution,proto3,enum=officeconvertapi.v1.ConversionResolution" json:"resolution,omitempty"`
unknownFields protoimpl.UnknownFields
sizeCache protoimpl.SizeCache
}
@@ -312,6 +388,13 @@ func (x *CreateConversionRequest) GetSourceFilename() string {
return ""
}
func (x *CreateConversionRequest) GetResolution() ConversionResolution {
if x != nil {
return x.Resolution
}
return ConversionResolution_CONVERSION_RESOLUTION_UNSPECIFIED
}
// CreateConversionResponse returns upload details for the session.
type CreateConversionResponse struct {
state protoimpl.MessageState `protogen:"open.v1"`
@@ -822,15 +905,20 @@ const file_officeconvertapi_v1_conversion_proto_rawDesc = "" +
"\x05index\x18\x01 \x01(\x05R\x05index\x12\x1f\n" +
"\vnotes_plain\x18\x02 \x01(\tR\n" +
"notesPlain\x12\x1b\n" +
"\timage_url\x18\x03 \x01(\tR\bimageUrl\"\xc8\x01\n" +
"\timage_url\x18\x03 \x01(\tR\bimageUrl\"\xf6\x01\n" +
"\tSlideDeck\x12#\n" +
"\rconversion_id\x18\x01 \x01(\tR\fconversionId\x12'\n" +
"\x0fsource_filename\x18\x02 \x01(\tR\x0esourceFilename\x122\n" +
"\x06slides\x18\x03 \x03(\v2\x1a.officeconvertapi.v1.SlideR\x06slides\x129\n" +
"\n" +
"created_at\x18\x04 \x01(\v2\x1a.google.protobuf.TimestampR\tcreatedAt\"B\n" +
"created_at\x18\x04 \x01(\v2\x1a.google.protobuf.TimestampR\tcreatedAt\x12\x14\n" +
"\x05width\x18\x05 \x01(\x05R\x05width\x12\x16\n" +
"\x06height\x18\x06 \x01(\x05R\x06height\"\x8d\x01\n" +
"\x17CreateConversionRequest\x12'\n" +
"\x0fsource_filename\x18\x01 \x01(\tR\x0esourceFilename\"\xea\x01\n" +
"\x0fsource_filename\x18\x01 \x01(\tR\x0esourceFilename\x12I\n" +
"\n" +
"resolution\x18\x02 \x01(\x0e2).officeconvertapi.v1.ConversionResolutionR\n" +
"resolution\"\xea\x01\n" +
"\x18CreateConversionResponse\x12#\n" +
"\rconversion_id\x18\x01 \x01(\tR\fconversionId\x12#\n" +
"\rupload_bucket\x18\x02 \x01(\tR\fuploadBucket\x12*\n" +
@@ -877,7 +965,14 @@ const file_officeconvertapi_v1_conversion_proto_rawDesc = "" +
"!CONVERSION_PHASE_EXTRACTING_NOTES\x10\x02\x12 \n" +
"\x1cCONVERSION_PHASE_PPTX_TO_PDF\x10\x03\x12\"\n" +
"\x1eCONVERSION_PHASE_PDF_TO_IMAGES\x10\x04\x12&\n" +
"\"CONVERSION_PHASE_UPLOADING_RESULTS\x10\x052\xcc\x04\n" +
"\"CONVERSION_PHASE_UPLOADING_RESULTS\x10\x05*\xd6\x01\n" +
"\x14ConversionResolution\x12%\n" +
"!CONVERSION_RESOLUTION_UNSPECIFIED\x10\x00\x12\x1c\n" +
"\x18CONVERSION_RESOLUTION_SD\x10\x01\x12\x1c\n" +
"\x18CONVERSION_RESOLUTION_HD\x10\x02\x12\x1d\n" +
"\x19CONVERSION_RESOLUTION_FHD\x10\x03\x12\x1d\n" +
"\x19CONVERSION_RESOLUTION_QHD\x10\x04\x12\x1d\n" +
"\x19CONVERSION_RESOLUTION_UHD\x10\x052\xcc\x04\n" +
"\x11ConversionService\x12q\n" +
"\x10CreateConversion\x12,.officeconvertapi.v1.CreateConversionRequest\x1a-.officeconvertapi.v1.CreateConversionResponse\"\x00\x12n\n" +
"\x0fStartConversion\x12+.officeconvertapi.v1.StartConversionRequest\x1a,.officeconvertapi.v1.StartConversionResponse\"\x00\x12z\n" +
@@ -897,49 +992,51 @@ func file_officeconvertapi_v1_conversion_proto_rawDescGZIP() []byte {
return file_officeconvertapi_v1_conversion_proto_rawDescData
}
var file_officeconvertapi_v1_conversion_proto_enumTypes = make([]protoimpl.EnumInfo, 2)
var file_officeconvertapi_v1_conversion_proto_enumTypes = make([]protoimpl.EnumInfo, 3)
var file_officeconvertapi_v1_conversion_proto_msgTypes = make([]protoimpl.MessageInfo, 12)
var file_officeconvertapi_v1_conversion_proto_goTypes = []any{
(ConversionStatus)(0), // 0: officeconvertapi.v1.ConversionStatus
(ConversionPhase)(0), // 1: officeconvertapi.v1.ConversionPhase
(*Slide)(nil), // 2: officeconvertapi.v1.Slide
(*SlideDeck)(nil), // 3: officeconvertapi.v1.SlideDeck
(*CreateConversionRequest)(nil), // 4: officeconvertapi.v1.CreateConversionRequest
(*CreateConversionResponse)(nil), // 5: officeconvertapi.v1.CreateConversionResponse
(*StartConversionRequest)(nil), // 6: officeconvertapi.v1.StartConversionRequest
(*StartConversionResponse)(nil), // 7: officeconvertapi.v1.StartConversionResponse
(*GetConversionStatusRequest)(nil), // 8: officeconvertapi.v1.GetConversionStatusRequest
(*GetConversionStatusResponse)(nil), // 9: officeconvertapi.v1.GetConversionStatusResponse
(*GetSlideDeckRequest)(nil), // 10: officeconvertapi.v1.GetSlideDeckRequest
(*GetSlideDeckResponse)(nil), // 11: officeconvertapi.v1.GetSlideDeckResponse
(*DeleteConversionRequest)(nil), // 12: officeconvertapi.v1.DeleteConversionRequest
(*DeleteConversionResponse)(nil), // 13: officeconvertapi.v1.DeleteConversionResponse
(*timestamppb.Timestamp)(nil), // 14: google.protobuf.Timestamp
(ConversionResolution)(0), // 2: officeconvertapi.v1.ConversionResolution
(*Slide)(nil), // 3: officeconvertapi.v1.Slide
(*SlideDeck)(nil), // 4: officeconvertapi.v1.SlideDeck
(*CreateConversionRequest)(nil), // 5: officeconvertapi.v1.CreateConversionRequest
(*CreateConversionResponse)(nil), // 6: officeconvertapi.v1.CreateConversionResponse
(*StartConversionRequest)(nil), // 7: officeconvertapi.v1.StartConversionRequest
(*StartConversionResponse)(nil), // 8: officeconvertapi.v1.StartConversionResponse
(*GetConversionStatusRequest)(nil), // 9: officeconvertapi.v1.GetConversionStatusRequest
(*GetConversionStatusResponse)(nil), // 10: officeconvertapi.v1.GetConversionStatusResponse
(*GetSlideDeckRequest)(nil), // 11: officeconvertapi.v1.GetSlideDeckRequest
(*GetSlideDeckResponse)(nil), // 12: officeconvertapi.v1.GetSlideDeckResponse
(*DeleteConversionRequest)(nil), // 13: officeconvertapi.v1.DeleteConversionRequest
(*DeleteConversionResponse)(nil), // 14: officeconvertapi.v1.DeleteConversionResponse
(*timestamppb.Timestamp)(nil), // 15: google.protobuf.Timestamp
}
var file_officeconvertapi_v1_conversion_proto_depIdxs = []int32{
2, // 0: officeconvertapi.v1.SlideDeck.slides:type_name -> officeconvertapi.v1.Slide
14, // 1: officeconvertapi.v1.SlideDeck.created_at:type_name -> google.protobuf.Timestamp
14, // 2: officeconvertapi.v1.CreateConversionResponse.expires_at:type_name -> google.protobuf.Timestamp
0, // 3: officeconvertapi.v1.StartConversionResponse.status:type_name -> officeconvertapi.v1.ConversionStatus
0, // 4: officeconvertapi.v1.GetConversionStatusResponse.status:type_name -> officeconvertapi.v1.ConversionStatus
14, // 5: officeconvertapi.v1.GetConversionStatusResponse.updated_at:type_name -> google.protobuf.Timestamp
1, // 6: officeconvertapi.v1.GetConversionStatusResponse.phase:type_name -> officeconvertapi.v1.ConversionPhase
3, // 7: officeconvertapi.v1.GetSlideDeckResponse.slide_deck:type_name -> officeconvertapi.v1.SlideDeck
4, // 8: officeconvertapi.v1.ConversionService.CreateConversion:input_type -> officeconvertapi.v1.CreateConversionRequest
6, // 9: officeconvertapi.v1.ConversionService.StartConversion:input_type -> officeconvertapi.v1.StartConversionRequest
8, // 10: officeconvertapi.v1.ConversionService.GetConversionStatus:input_type -> officeconvertapi.v1.GetConversionStatusRequest
10, // 11: officeconvertapi.v1.ConversionService.GetSlideDeck:input_type -> officeconvertapi.v1.GetSlideDeckRequest
12, // 12: officeconvertapi.v1.ConversionService.DeleteConversion:input_type -> officeconvertapi.v1.DeleteConversionRequest
5, // 13: officeconvertapi.v1.ConversionService.CreateConversion:output_type -> officeconvertapi.v1.CreateConversionResponse
7, // 14: officeconvertapi.v1.ConversionService.StartConversion:output_type -> officeconvertapi.v1.StartConversionResponse
9, // 15: officeconvertapi.v1.ConversionService.GetConversionStatus:output_type -> officeconvertapi.v1.GetConversionStatusResponse
11, // 16: officeconvertapi.v1.ConversionService.GetSlideDeck:output_type -> officeconvertapi.v1.GetSlideDeckResponse
13, // 17: officeconvertapi.v1.ConversionService.DeleteConversion:output_type -> officeconvertapi.v1.DeleteConversionResponse
13, // [13:18] is the sub-list for method output_type
8, // [8:13] is the sub-list for method input_type
8, // [8:8] is the sub-list for extension type_name
8, // [8:8] is the sub-list for extension extendee
0, // [0:8] is the sub-list for field type_name
3, // 0: officeconvertapi.v1.SlideDeck.slides:type_name -> officeconvertapi.v1.Slide
15, // 1: officeconvertapi.v1.SlideDeck.created_at:type_name -> google.protobuf.Timestamp
2, // 2: officeconvertapi.v1.CreateConversionRequest.resolution:type_name -> officeconvertapi.v1.ConversionResolution
15, // 3: officeconvertapi.v1.CreateConversionResponse.expires_at:type_name -> google.protobuf.Timestamp
0, // 4: officeconvertapi.v1.StartConversionResponse.status:type_name -> officeconvertapi.v1.ConversionStatus
0, // 5: officeconvertapi.v1.GetConversionStatusResponse.status:type_name -> officeconvertapi.v1.ConversionStatus
15, // 6: officeconvertapi.v1.GetConversionStatusResponse.updated_at:type_name -> google.protobuf.Timestamp
1, // 7: officeconvertapi.v1.GetConversionStatusResponse.phase:type_name -> officeconvertapi.v1.ConversionPhase
4, // 8: officeconvertapi.v1.GetSlideDeckResponse.slide_deck:type_name -> officeconvertapi.v1.SlideDeck
5, // 9: officeconvertapi.v1.ConversionService.CreateConversion:input_type -> officeconvertapi.v1.CreateConversionRequest
7, // 10: officeconvertapi.v1.ConversionService.StartConversion:input_type -> officeconvertapi.v1.StartConversionRequest
9, // 11: officeconvertapi.v1.ConversionService.GetConversionStatus:input_type -> officeconvertapi.v1.GetConversionStatusRequest
11, // 12: officeconvertapi.v1.ConversionService.GetSlideDeck:input_type -> officeconvertapi.v1.GetSlideDeckRequest
13, // 13: officeconvertapi.v1.ConversionService.DeleteConversion:input_type -> officeconvertapi.v1.DeleteConversionRequest
6, // 14: officeconvertapi.v1.ConversionService.CreateConversion:output_type -> officeconvertapi.v1.CreateConversionResponse
8, // 15: officeconvertapi.v1.ConversionService.StartConversion:output_type -> officeconvertapi.v1.StartConversionResponse
10, // 16: officeconvertapi.v1.ConversionService.GetConversionStatus:output_type -> officeconvertapi.v1.GetConversionStatusResponse
12, // 17: officeconvertapi.v1.ConversionService.GetSlideDeck:output_type -> officeconvertapi.v1.GetSlideDeckResponse
14, // 18: officeconvertapi.v1.ConversionService.DeleteConversion:output_type -> officeconvertapi.v1.DeleteConversionResponse
14, // [14:19] is the sub-list for method output_type
9, // [9:14] is the sub-list for method input_type
9, // [9:9] is the sub-list for extension type_name
9, // [9:9] is the sub-list for extension extendee
0, // [0:9] is the sub-list for field type_name
}
func init() { file_officeconvertapi_v1_conversion_proto_init() }
@@ -952,7 +1049,7 @@ func file_officeconvertapi_v1_conversion_proto_init() {
File: protoimpl.DescBuilder{
GoPackagePath: reflect.TypeOf(x{}).PkgPath(),
RawDescriptor: unsafe.Slice(unsafe.StringData(file_officeconvertapi_v1_conversion_proto_rawDesc), len(file_officeconvertapi_v1_conversion_proto_rawDesc)),
NumEnums: 2,
NumEnums: 3,
NumMessages: 12,
NumExtensions: 0,
NumServices: 1,
gen/python/officeconvertapi/v1/conversion_pb2.py
+30 -28
View File
@@ -25,7 +25,7 @@ _sym_db = _symbol_database.Default()
from google.protobuf import timestamp_pb2 as google_dot_protobuf_dot_timestamp__pb2
DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n$officeconvertapi/v1/conversion.proto\x12\x13officeconvertapi.v1\x1a\x1fgoogle/protobuf/timestamp.proto\"[\n\x05Slide\x12\x14\n\x05index\x18\x01 \x01(\x05R\x05index\x12\x1f\n\x0bnotes_plain\x18\x02 \x01(\tR\nnotesPlain\x12\x1b\n\timage_url\x18\x03 \x01(\tR\x08imageUrl\"\xc8\x01\n\tSlideDeck\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12\'\n\x0fsource_filename\x18\x02 \x01(\tR\x0esourceFilename\x12\x32\n\x06slides\x18\x03 \x03(\x0b\x32\x1a.officeconvertapi.v1.SlideR\x06slides\x12\x39\n\ncreated_at\x18\x04 \x01(\x0b\x32\x1a.google.protobuf.TimestampR\tcreatedAt\"B\n\x17\x43reateConversionRequest\x12\'\n\x0fsource_filename\x18\x01 \x01(\tR\x0esourceFilename\"\xea\x01\n\x18\x43reateConversionResponse\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12#\n\rupload_bucket\x18\x02 \x01(\tR\x0cuploadBucket\x12*\n\x11upload_object_key\x18\x03 \x01(\tR\x0fuploadObjectKey\x12\x1d\n\nupload_url\x18\x04 \x01(\tR\tuploadUrl\x12\x39\n\nexpires_at\x18\x05 \x01(\x0b\x32\x1a.google.protobuf.TimestampR\texpiresAt\"=\n\x16StartConversionRequest\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\"}\n\x17StartConversionResponse\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12=\n\x06status\x18\x02 \x01(\x0e\x32%.officeconvertapi.v1.ConversionStatusR\x06status\"A\n\x1aGetConversionStatusRequest\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\"\xeb\x02\n\x1bGetConversionStatusResponse\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12=\n\x06status\x18\x02 \x01(\x0e\x32%.officeconvertapi.v1.ConversionStatusR\x06status\x12#\n\rerror_message\x18\x03 \x01(\tR\x0c\x65rrorMessage\x12\x39\n\nupdated_at\x18\x04 \x01(\x0b\x32\x1a.google.protobuf.TimestampR\tupdatedAt\x12:\n\x05phase\x18\x05 \x01(\x0e\x32$.officeconvertapi.v1.ConversionPhaseR\x05phase\x12)\n\x10\x63urrent_progress\x18\x06 \x01(\x05R\x0f\x63urrentProgress\x12!\n\x0cmax_progress\x18\x07 \x01(\x05R\x0bmaxProgress\":\n\x13GetSlideDeckRequest\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\"U\n\x14GetSlideDeckResponse\x12=\n\nslide_deck\x18\x01 \x01(\x0b\x32\x1e.officeconvertapi.v1.SlideDeckR\tslideDeck\">\n\x17\x44\x65leteConversionRequest\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\"Y\n\x18\x44\x65leteConversionResponse\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12\x18\n\x07\x64\x65leted\x18\x02 \x01(\x08R\x07\x64\x65leted*\xb2\x01\n\x10\x43onversionStatus\x12!\n\x1d\x43ONVERSION_STATUS_UNSPECIFIED\x10\x00\x12\x1d\n\x19\x43ONVERSION_STATUS_PENDING\x10\x01\x12\x1d\n\x19\x43ONVERSION_STATUS_RUNNING\x10\x02\x12\x1f\n\x1b\x43ONVERSION_STATUS_SUCCEEDED\x10\x03\x12\x1c\n\x18\x43ONVERSION_STATUS_FAILED\x10\x04*\xe7\x01\n\x0f\x43onversionPhase\x12 \n\x1c\x43ONVERSION_PHASE_UNSPECIFIED\x10\x00\x12\x1d\n\x19\x43ONVERSION_PHASE_INACTIVE\x10\x01\x12%\n!CONVERSION_PHASE_EXTRACTING_NOTES\x10\x02\x12 \n\x1c\x43ONVERSION_PHASE_PPTX_TO_PDF\x10\x03\x12\"\n\x1e\x43ONVERSION_PHASE_PDF_TO_IMAGES\x10\x04\x12&\n\"CONVERSION_PHASE_UPLOADING_RESULTS\x10\x05\x32\xcc\x04\n\x11\x43onversionService\x12q\n\x10\x43reateConversion\x12,.officeconvertapi.v1.CreateConversionRequest\x1a-.officeconvertapi.v1.CreateConversionResponse\"\x00\x12n\n\x0fStartConversion\x12+.officeconvertapi.v1.StartConversionRequest\x1a,.officeconvertapi.v1.StartConversionResponse\"\x00\x12z\n\x13GetConversionStatus\x12/.officeconvertapi.v1.GetConversionStatusRequest\x1a\x30.officeconvertapi.v1.GetConversionStatusResponse\"\x00\x12\x65\n\x0cGetSlideDeck\x12(.officeconvertapi.v1.GetSlideDeckRequest\x1a).officeconvertapi.v1.GetSlideDeckResponse\"\x00\x12q\n\x10\x44\x65leteConversion\x12,.officeconvertapi.v1.DeleteConversionRequest\x1a-.officeconvertapi.v1.DeleteConversionResponse\"\x00\x42LZJgithub.com/end/officeconvert/gen/go/officeconvertapi/v1;officeconvertapiv1b\x06proto3')
DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n$officeconvertapi/v1/conversion.proto\x12\x13officeconvertapi.v1\x1a\x1fgoogle/protobuf/timestamp.proto\"[\n\x05Slide\x12\x14\n\x05index\x18\x01 \x01(\x05R\x05index\x12\x1f\n\x0bnotes_plain\x18\x02 \x01(\tR\nnotesPlain\x12\x1b\n\timage_url\x18\x03 \x01(\tR\x08imageUrl\"\xf6\x01\n\tSlideDeck\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12\'\n\x0fsource_filename\x18\x02 \x01(\tR\x0esourceFilename\x12\x32\n\x06slides\x18\x03 \x03(\x0b\x32\x1a.officeconvertapi.v1.SlideR\x06slides\x12\x39\n\ncreated_at\x18\x04 \x01(\x0b\x32\x1a.google.protobuf.TimestampR\tcreatedAt\x12\x14\n\x05width\x18\x05 \x01(\x05R\x05width\x12\x16\n\x06height\x18\x06 \x01(\x05R\x06height\"\x8d\x01\n\x17\x43reateConversionRequest\x12\'\n\x0fsource_filename\x18\x01 \x01(\tR\x0esourceFilename\x12I\n\nresolution\x18\x02 \x01(\x0e\x32).officeconvertapi.v1.ConversionResolutionR\nresolution\"\xea\x01\n\x18\x43reateConversionResponse\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12#\n\rupload_bucket\x18\x02 \x01(\tR\x0cuploadBucket\x12*\n\x11upload_object_key\x18\x03 \x01(\tR\x0fuploadObjectKey\x12\x1d\n\nupload_url\x18\x04 \x01(\tR\tuploadUrl\x12\x39\n\nexpires_at\x18\x05 \x01(\x0b\x32\x1a.google.protobuf.TimestampR\texpiresAt\"=\n\x16StartConversionRequest\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\"}\n\x17StartConversionResponse\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12=\n\x06status\x18\x02 \x01(\x0e\x32%.officeconvertapi.v1.ConversionStatusR\x06status\"A\n\x1aGetConversionStatusRequest\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\"\xeb\x02\n\x1bGetConversionStatusResponse\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12=\n\x06status\x18\x02 \x01(\x0e\x32%.officeconvertapi.v1.ConversionStatusR\x06status\x12#\n\rerror_message\x18\x03 \x01(\tR\x0c\x65rrorMessage\x12\x39\n\nupdated_at\x18\x04 \x01(\x0b\x32\x1a.google.protobuf.TimestampR\tupdatedAt\x12:\n\x05phase\x18\x05 \x01(\x0e\x32$.officeconvertapi.v1.ConversionPhaseR\x05phase\x12)\n\x10\x63urrent_progress\x18\x06 \x01(\x05R\x0f\x63urrentProgress\x12!\n\x0cmax_progress\x18\x07 \x01(\x05R\x0bmaxProgress\":\n\x13GetSlideDeckRequest\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\"U\n\x14GetSlideDeckResponse\x12=\n\nslide_deck\x18\x01 \x01(\x0b\x32\x1e.officeconvertapi.v1.SlideDeckR\tslideDeck\">\n\x17\x44\x65leteConversionRequest\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\"Y\n\x18\x44\x65leteConversionResponse\x12#\n\rconversion_id\x18\x01 \x01(\tR\x0c\x63onversionId\x12\x18\n\x07\x64\x65leted\x18\x02 \x01(\x08R\x07\x64\x65leted*\xb2\x01\n\x10\x43onversionStatus\x12!\n\x1d\x43ONVERSION_STATUS_UNSPECIFIED\x10\x00\x12\x1d\n\x19\x43ONVERSION_STATUS_PENDING\x10\x01\x12\x1d\n\x19\x43ONVERSION_STATUS_RUNNING\x10\x02\x12\x1f\n\x1b\x43ONVERSION_STATUS_SUCCEEDED\x10\x03\x12\x1c\n\x18\x43ONVERSION_STATUS_FAILED\x10\x04*\xe7\x01\n\x0f\x43onversionPhase\x12 \n\x1c\x43ONVERSION_PHASE_UNSPECIFIED\x10\x00\x12\x1d\n\x19\x43ONVERSION_PHASE_INACTIVE\x10\x01\x12%\n!CONVERSION_PHASE_EXTRACTING_NOTES\x10\x02\x12 \n\x1c\x43ONVERSION_PHASE_PPTX_TO_PDF\x10\x03\x12\"\n\x1e\x43ONVERSION_PHASE_PDF_TO_IMAGES\x10\x04\x12&\n\"CONVERSION_PHASE_UPLOADING_RESULTS\x10\x05*\xd6\x01\n\x14\x43onversionResolution\x12%\n!CONVERSION_RESOLUTION_UNSPECIFIED\x10\x00\x12\x1c\n\x18\x43ONVERSION_RESOLUTION_SD\x10\x01\x12\x1c\n\x18\x43ONVERSION_RESOLUTION_HD\x10\x02\x12\x1d\n\x19\x43ONVERSION_RESOLUTION_FHD\x10\x03\x12\x1d\n\x19\x43ONVERSION_RESOLUTION_QHD\x10\x04\x12\x1d\n\x19\x43ONVERSION_RESOLUTION_UHD\x10\x05\x32\xcc\x04\n\x11\x43onversionService\x12q\n\x10\x43reateConversion\x12,.officeconvertapi.v1.CreateConversionRequest\x1a-.officeconvertapi.v1.CreateConversionResponse\"\x00\x12n\n\x0fStartConversion\x12+.officeconvertapi.v1.StartConversionRequest\x1a,.officeconvertapi.v1.StartConversionResponse\"\x00\x12z\n\x13GetConversionStatus\x12/.officeconvertapi.v1.GetConversionStatusRequest\x1a\x30.officeconvertapi.v1.GetConversionStatusResponse\"\x00\x12\x65\n\x0cGetSlideDeck\x12(.officeconvertapi.v1.GetSlideDeckRequest\x1a).officeconvertapi.v1.GetSlideDeckResponse\"\x00\x12q\n\x10\x44\x65leteConversion\x12,.officeconvertapi.v1.DeleteConversionRequest\x1a-.officeconvertapi.v1.DeleteConversionResponse\"\x00\x42LZJgithub.com/end/officeconvert/gen/go/officeconvertapi/v1;officeconvertapiv1b\x06proto3')
_globals = globals()
_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
@@ -33,34 +33,36 @@ _builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, 'officeconvertapi.v1.convers
if not _descriptor._USE_C_DESCRIPTORS:
_globals['DESCRIPTOR']._loaded_options = None
_globals['DESCRIPTOR']._serialized_options = b'ZJgithub.com/end/officeconvert/gen/go/officeconvertapi/v1;officeconvertapiv1'
_globals['_CONVERSIONSTATUS']._serialized_start=1621
_globals['_CONVERSIONSTATUS']._serialized_end=1799
_globals['_CONVERSIONPHASE']._serialized_start=1802
_globals['_CONVERSIONPHASE']._serialized_end=2033
_globals['_CONVERSIONSTATUS']._serialized_start=1743
_globals['_CONVERSIONSTATUS']._serialized_end=1921
_globals['_CONVERSIONPHASE']._serialized_start=1924
_globals['_CONVERSIONPHASE']._serialized_end=2155
_globals['_CONVERSIONRESOLUTION']._serialized_start=2158
_globals['_CONVERSIONRESOLUTION']._serialized_end=2372
_globals['_SLIDE']._serialized_start=94
_globals['_SLIDE']._serialized_end=185
_globals['_SLIDEDECK']._serialized_start=188
_globals['_SLIDEDECK']._serialized_end=388
_globals['_CREATECONVERSIONREQUEST']._serialized_start=390
_globals['_CREATECONVERSIONREQUEST']._serialized_end=456
_globals['_CREATECONVERSIONRESPONSE']._serialized_start=459
_globals['_CREATECONVERSIONRESPONSE']._serialized_end=693
_globals['_STARTCONVERSIONREQUEST']._serialized_start=695
_globals['_STARTCONVERSIONREQUEST']._serialized_end=756
_globals['_STARTCONVERSIONRESPONSE']._serialized_start=758
_globals['_STARTCONVERSIONRESPONSE']._serialized_end=883
_globals['_GETCONVERSIONSTATUSREQUEST']._serialized_start=885
_globals['_GETCONVERSIONSTATUSREQUEST']._serialized_end=950
_globals['_GETCONVERSIONSTATUSRESPONSE']._serialized_start=953
_globals['_GETCONVERSIONSTATUSRESPONSE']._serialized_end=1316
_globals['_GETSLIDEDECKREQUEST']._serialized_start=1318
_globals['_GETSLIDEDECKREQUEST']._serialized_end=1376
_globals['_GETSLIDEDECKRESPONSE']._serialized_start=1378
_globals['_GETSLIDEDECKRESPONSE']._serialized_end=1463
_globals['_DELETECONVERSIONREQUEST']._serialized_start=1465
_globals['_DELETECONVERSIONREQUEST']._serialized_end=1527
_globals['_DELETECONVERSIONRESPONSE']._serialized_start=1529
_globals['_DELETECONVERSIONRESPONSE']._serialized_end=1618
_globals['_CONVERSIONSERVICE']._serialized_start=2036
_globals['_CONVERSIONSERVICE']._serialized_end=2624
_globals['_SLIDEDECK']._serialized_end=434
_globals['_CREATECONVERSIONREQUEST']._serialized_start=437
_globals['_CREATECONVERSIONREQUEST']._serialized_end=578
_globals['_CREATECONVERSIONRESPONSE']._serialized_start=581
_globals['_CREATECONVERSIONRESPONSE']._serialized_end=815
_globals['_STARTCONVERSIONREQUEST']._serialized_start=817
_globals['_STARTCONVERSIONREQUEST']._serialized_end=878
_globals['_STARTCONVERSIONRESPONSE']._serialized_start=880
_globals['_STARTCONVERSIONRESPONSE']._serialized_end=1005
_globals['_GETCONVERSIONSTATUSREQUEST']._serialized_start=1007
_globals['_GETCONVERSIONSTATUSREQUEST']._serialized_end=1072
_globals['_GETCONVERSIONSTATUSRESPONSE']._serialized_start=1075
_globals['_GETCONVERSIONSTATUSRESPONSE']._serialized_end=1438
_globals['_GETSLIDEDECKREQUEST']._serialized_start=1440
_globals['_GETSLIDEDECKREQUEST']._serialized_end=1498
_globals['_GETSLIDEDECKRESPONSE']._serialized_start=1500
_globals['_GETSLIDEDECKRESPONSE']._serialized_end=1585
_globals['_DELETECONVERSIONREQUEST']._serialized_start=1587
_globals['_DELETECONVERSIONREQUEST']._serialized_end=1649
_globals['_DELETECONVERSIONRESPONSE']._serialized_start=1651
_globals['_DELETECONVERSIONRESPONSE']._serialized_end=1740
_globals['_CONVERSIONSERVICE']._serialized_start=2375
_globals['_CONVERSIONSERVICE']._serialized_end=2963
# @@protoc_insertion_point(module_scope)
gen/python/officeconvertapi/v1/conversion_pb2.pyi
+25 -4
View File
@@ -26,6 +26,15 @@ class ConversionPhase(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
CONVERSION_PHASE_PPTX_TO_PDF: _ClassVar[ConversionPhase]
CONVERSION_PHASE_PDF_TO_IMAGES: _ClassVar[ConversionPhase]
CONVERSION_PHASE_UPLOADING_RESULTS: _ClassVar[ConversionPhase]
class ConversionResolution(int, metaclass=_enum_type_wrapper.EnumTypeWrapper):
__slots__ = ()
CONVERSION_RESOLUTION_UNSPECIFIED: _ClassVar[ConversionResolution]
CONVERSION_RESOLUTION_SD: _ClassVar[ConversionResolution]
CONVERSION_RESOLUTION_HD: _ClassVar[ConversionResolution]
CONVERSION_RESOLUTION_FHD: _ClassVar[ConversionResolution]
CONVERSION_RESOLUTION_QHD: _ClassVar[ConversionResolution]
CONVERSION_RESOLUTION_UHD: _ClassVar[ConversionResolution]
CONVERSION_STATUS_UNSPECIFIED: ConversionStatus
CONVERSION_STATUS_PENDING: ConversionStatus
CONVERSION_STATUS_RUNNING: ConversionStatus
@@ -37,6 +46,12 @@ CONVERSION_PHASE_EXTRACTING_NOTES: ConversionPhase
CONVERSION_PHASE_PPTX_TO_PDF: ConversionPhase
CONVERSION_PHASE_PDF_TO_IMAGES: ConversionPhase
CONVERSION_PHASE_UPLOADING_RESULTS: ConversionPhase
CONVERSION_RESOLUTION_UNSPECIFIED: ConversionResolution
CONVERSION_RESOLUTION_SD: ConversionResolution
CONVERSION_RESOLUTION_HD: ConversionResolution
CONVERSION_RESOLUTION_FHD: ConversionResolution
CONVERSION_RESOLUTION_QHD: ConversionResolution
CONVERSION_RESOLUTION_UHD: ConversionResolution
class Slide(_message.Message):
__slots__ = ("index", "notes_plain", "image_url")
@@ -49,22 +64,28 @@ class Slide(_message.Message):
def __init__(self, index: _Optional[int] = ..., notes_plain: _Optional[str] = ..., image_url: _Optional[str] = ...) -> None: ...
class SlideDeck(_message.Message):
__slots__ = ("conversion_id", "source_filename", "slides", "created_at")
__slots__ = ("conversion_id", "source_filename", "slides", "created_at", "width", "height")
CONVERSION_ID_FIELD_NUMBER: _ClassVar[int]
SOURCE_FILENAME_FIELD_NUMBER: _ClassVar[int]
SLIDES_FIELD_NUMBER: _ClassVar[int]
CREATED_AT_FIELD_NUMBER: _ClassVar[int]
WIDTH_FIELD_NUMBER: _ClassVar[int]
HEIGHT_FIELD_NUMBER: _ClassVar[int]
conversion_id: str
source_filename: str
slides: _containers.RepeatedCompositeFieldContainer[Slide]
created_at: _timestamp_pb2.Timestamp
def __init__(self, conversion_id: _Optional[str] = ..., source_filename: _Optional[str] = ..., slides: _Optional[_Iterable[_Union[Slide, _Mapping]]] = ..., created_at: _Optional[_Union[datetime.datetime, _timestamp_pb2.Timestamp, _Mapping]] = ...) -> None: ...
width: int
height: int
def __init__(self, conversion_id: _Optional[str] = ..., source_filename: _Optional[str] = ..., slides: _Optional[_Iterable[_Union[Slide, _Mapping]]] = ..., created_at: _Optional[_Union[datetime.datetime, _timestamp_pb2.Timestamp, _Mapping]] = ..., width: _Optional[int] = ..., height: _Optional[int] = ...) -> None: ...
class CreateConversionRequest(_message.Message):
__slots__ = ("source_filename",)
__slots__ = ("source_filename", "resolution")
SOURCE_FILENAME_FIELD_NUMBER: _ClassVar[int]
RESOLUTION_FIELD_NUMBER: _ClassVar[int]
source_filename: str
def __init__(self, source_filename: _Optional[str] = ...) -> None: ...
resolution: ConversionResolution
def __init__(self, source_filename: _Optional[str] = ..., resolution: _Optional[_Union[ConversionResolution, str]] = ...) -> None: ...
class CreateConversionResponse(_message.Message):
__slots__ = ("conversion_id", "upload_bucket", "upload_object_key", "upload_url", "expires_at")
proto/officeconvertapi/v1/conversion.proto
+13
View File
@@ -43,6 +43,16 @@ enum ConversionPhase {
CONVERSION_PHASE_UPLOADING_RESULTS = 5;
}
// ConversionResolution represents preset output quality targets.
enum ConversionResolution {
CONVERSION_RESOLUTION_UNSPECIFIED = 0;
CONVERSION_RESOLUTION_SD = 1;
CONVERSION_RESOLUTION_HD = 2;
CONVERSION_RESOLUTION_FHD = 3;
CONVERSION_RESOLUTION_QHD = 4;
CONVERSION_RESOLUTION_UHD = 5;
}
// Slide contains extracted notes and the rendered image URL for one slide.
message Slide {
int32 index = 1;
@@ -56,11 +66,14 @@ message SlideDeck {
string source_filename = 2;
repeated Slide slides = 3;
google.protobuf.Timestamp created_at = 4;
int32 width = 5;
int32 height = 6;
}
// CreateConversionRequest starts a conversion session.
message CreateConversionRequest {
string source_filename = 1;
ConversionResolution resolution = 2;
}
// CreateConversionResponse returns upload details for the session.
python/packages/officeconvert/src/officeconvert/conversion.py
+136 -8
View File
@@ -5,6 +5,7 @@ from __future__ import annotations
from collections.abc import Callable
from dataclasses import dataclass
import logging
import math
from pathlib import Path
import subprocess
from typing import Iterable
@@ -27,6 +28,12 @@ class SlideDeckResult:
source_filename: str
slides: list[SlideArtifact]
width: int
height: int
inferred_dpi: int
pptx_to_pdf_timeout_s: int
pdf_to_images_timeout_s: int
pdf_to_images_page_timeout_s: int
ProgressCallback = Callable[[str, int, int], None]
@@ -36,6 +43,21 @@ PHASE_EXTRACTING_NOTES = "extracting_notes"
PHASE_PPTX_TO_PDF = "pptx_to_pdf"
PHASE_PDF_TO_IMAGES = "pdf_to_images"
RESOLUTION_SD = "sd"
RESOLUTION_HD = "hd"
RESOLUTION_FHD = "fhd"
RESOLUTION_QHD = "qhd"
RESOLUTION_UHD = "uhd"
_SHORT_EDGE_PIXELS_BY_RESOLUTION = {
RESOLUTION_SD: 480,
RESOLUTION_HD: 720,
RESOLUTION_FHD: 1080,
RESOLUTION_QHD: 1440,
RESOLUTION_UHD: 2160,
}
_EMU_PER_INCH = 914400
logger = logging.getLogger("uvicorn.error")
@@ -108,6 +130,8 @@ def render_pdf_to_images(
out_dir: Path,
*,
dpi: int = 72,
target_width: int | None = None,
target_height: int | None = None,
image_format: str = "png",
timeout_s: int = 120,
total_pages: int | None = None,
@@ -132,14 +156,24 @@ def render_pdf_to_images(
"""
if not pdf_path.exists():
raise FileNotFoundError(f"source PDF does not exist: {pdf_path}")
if (target_width is None) != (target_height is None):
raise ValueError("target_width and target_height must be provided together")
if target_width is not None and target_width <= 0:
raise ValueError("target_width must be greater than zero")
if target_height is not None and target_height <= 0:
raise ValueError("target_height must be greater than zero")
out_dir.mkdir(parents=True, exist_ok=True)
scale_args: list[str] = []
if target_width is not None and target_height is not None:
scale_args = ["-scale-to-x", str(target_width), "-scale-to-y", str(target_height)]
if total_pages is None:
prefix_path = out_dir / "slide"
command = [
"pdftoppm",
"-r",
str(dpi),
*scale_args,
f"-{image_format}",
str(pdf_path.resolve()),
str(prefix_path),
@@ -156,7 +190,7 @@ def render_pdf_to_images(
message = (
"Poppler rasterization timed out after "
f"{timeout_s} seconds while rendering {pdf_path.name}; "
"increase conversion PDF render timeout cap or lower image DPI"
"increase conversion PDF render timeout cap or lower output resolution"
)
logger.error(message, exc_info=True)
raise ConversionTimeoutError(message) from exc
@@ -175,6 +209,7 @@ def render_pdf_to_images(
"pdftoppm",
"-r",
str(dpi),
*scale_args,
f"-{image_format}",
"-f",
str(page_index),
@@ -202,7 +237,7 @@ def render_pdf_to_images(
message = (
"Poppler rasterization timed out while rendering page "
f"{page_index}/{total_pages} of {pdf_path.name}; "
f"{timeout_context}. Increase timeout settings or lower image DPI."
f"{timeout_context}. Increase timeout settings or lower output resolution."
)
logger.error(message, exc_info=True)
raise ConversionTimeoutError(message) from exc
@@ -254,7 +289,7 @@ def convert_pptx_to_slidedeck(
pptx_path: Path,
work_dir: Path,
*,
dpi: int = 72,
resolution: str = RESOLUTION_FHD,
image_format: str = "png",
pptx_to_pdf_timeout_s: int = 180,
pdf_to_images_timeout_s: int = 1800,
@@ -273,7 +308,7 @@ def convert_pptx_to_slidedeck(
Args:
pptx_path: Source `.pptx` path.
work_dir: Scratch directory for generated outputs.
dpi: Rasterization DPI for output slide images.
resolution: Output resolution preset (`sd`, `hd`, `fhd`, `qhd`, `uhd`).
image_format: Output image format accepted by `pdftoppm`.
pptx_to_pdf_timeout_s: Timeout in seconds for the LibreOffice subprocess.
pdf_to_images_timeout_s: Timeout in seconds for the Poppler subprocess.
@@ -292,6 +327,18 @@ def convert_pptx_to_slidedeck(
_emit_progress(progress_callback, PHASE_EXTRACTING_NOTES, 0, 1)
notes = extract_slide_notes(pptx_path)
_emit_progress(progress_callback, PHASE_EXTRACTING_NOTES, 1, 1)
slide_width, slide_height = _read_slide_size_emu(pptx_path)
output_width, output_height = _infer_output_dimensions_from_slide_size(
slide_width=slide_width,
slide_height=slide_height,
resolution=resolution,
)
inferred_dpi = infer_minimum_raster_dpi(
slide_width_emu=slide_width,
slide_height_emu=slide_height,
output_width_px=output_width,
output_height_px=output_height,
)
slide_count = len(notes)
pptx_to_pdf_timeout = _compute_adaptive_timeout(
slide_count=slide_count,
@@ -317,12 +364,16 @@ def convert_pptx_to_slidedeck(
base_timeout_s=pdf_to_images_base_timeout_s,
)
logger.info(
"Conversion plan source=%s slides=%d dpi=%d image_format=%s "
"Conversion plan source=%s slides=%d inferred_dpi=%d image_format=%s "
"resolution=%s output_size=%dx%d "
"computed_timeouts_s[pptx_to_pdf_total=%d,pdf_to_images_total=%d,pdf_to_images_per_page=%d]",
pptx_path.name,
slide_count,
dpi,
inferred_dpi,
image_format,
resolution,
output_width,
output_height,
pptx_to_pdf_timeout,
pdf_to_images_timeout,
pdf_to_images_page_timeout,
@@ -330,7 +381,9 @@ def convert_pptx_to_slidedeck(
image_paths = render_pdf_to_images(
pdf_path,
image_dir,
dpi=dpi,
dpi=inferred_dpi,
target_width=output_width,
target_height=output_height,
image_format=image_format,
timeout_s=pdf_to_images_page_timeout,
total_pages=slide_count,
@@ -353,7 +406,82 @@ def convert_pptx_to_slidedeck(
SlideArtifact(index=index, image_path=image_path, notes_plain=note)
for index, (image_path, note) in enumerate(zip(image_paths, notes), start=1)
]
return SlideDeckResult(source_filename=pptx_path.name, slides=slides)
return SlideDeckResult(
source_filename=pptx_path.name,
slides=slides,
width=output_width,
height=output_height,
inferred_dpi=inferred_dpi,
pptx_to_pdf_timeout_s=pptx_to_pdf_timeout,
pdf_to_images_timeout_s=pdf_to_images_timeout,
pdf_to_images_page_timeout_s=pdf_to_images_page_timeout,
)
def infer_output_dimensions_for_resolution(
pptx_path: Path,
*,
resolution: str,
) -> tuple[int, int]:
"""Infer output image dimensions from source slide aspect ratio and preset."""
slide_width, slide_height = _read_slide_size_emu(pptx_path)
return _infer_output_dimensions_from_slide_size(
slide_width=slide_width,
slide_height=slide_height,
resolution=resolution,
)
def infer_minimum_raster_dpi(
*,
slide_width_emu: int,
slide_height_emu: int,
output_width_px: int,
output_height_px: int,
) -> int:
"""Compute the minimum DPI needed to reach target output dimensions."""
if slide_width_emu <= 0 or slide_height_emu <= 0:
raise ValueError("source slide dimensions must be greater than zero")
if output_width_px <= 0 or output_height_px <= 0:
raise ValueError("output dimensions must be greater than zero")
dpi_for_width = (output_width_px * _EMU_PER_INCH) / slide_width_emu
dpi_for_height = (output_height_px * _EMU_PER_INCH) / slide_height_emu
return max(1, math.ceil(max(dpi_for_width, dpi_for_height)))
def _read_slide_size_emu(pptx_path: Path) -> tuple[int, int]:
"""Read presentation slide size in English Metric Units (EMU)."""
if not pptx_path.exists():
raise FileNotFoundError(f"source PPTX does not exist: {pptx_path}")
presentation = Presentation(str(pptx_path.resolve()))
# Canonical python-pptx API on `Presentation` object.
slide_width = presentation.slide_width
slide_height = presentation.slide_height
if slide_width is None or slide_height is None:
raise ValueError("source presentation did not define slide dimensions")
slide_width = int(slide_width)
slide_height = int(slide_height)
if slide_width <= 0 or slide_height <= 0:
raise ValueError("source slide dimensions must be greater than zero")
return slide_width, slide_height
def _infer_output_dimensions_from_slide_size(
*,
slide_width: int,
slide_height: int,
resolution: str,
) -> tuple[int, int]:
"""Infer output dimensions from slide size and short-edge preset."""
normalized = resolution.strip().lower()
short_edge_pixels = _SHORT_EDGE_PIXELS_BY_RESOLUTION.get(normalized)
if short_edge_pixels is None:
raise ValueError(f"unsupported resolution preset: {resolution}")
if slide_width >= slide_height:
long_edge = max(1, round(short_edge_pixels * (slide_width / slide_height)))
return long_edge, short_edge_pixels
long_edge = max(1, round(short_edge_pixels * (slide_height / slide_width)))
return short_edge_pixels, long_edge
def _compute_adaptive_timeout(
python/packages/server/src/officeconvert_server/config.py
-2
View File
@@ -16,7 +16,6 @@ class ServerConfig:
s3_secure: bool
s3_public_endpoint: str
s3_session_ttl_seconds: int
conversion_image_dpi: int
conversion_pptx_to_pdf_timeout_seconds: int
conversion_pdf_to_images_timeout_seconds: int
conversion_pptx_to_pdf_base_timeout_seconds: int
@@ -35,7 +34,6 @@ def load_server_config() -> ServerConfig:
s3_secure=os.getenv("S3_USE_SSL", "false").lower() == "true",
s3_public_endpoint=os.getenv("S3_PUBLIC_ENDPOINT", "localhost:8333"),
s3_session_ttl_seconds=int(os.getenv("S3_SESSION_TTL_SECONDS", "3600")),
conversion_image_dpi=int(os.getenv("CONVERSION_IMAGE_DPI", "72")),
conversion_pptx_to_pdf_timeout_seconds=int(
os.getenv("CONVERSION_PPTX_TO_PDF_TIMEOUT_SECONDS", "180")
),
python/packages/server/src/officeconvert_server/models.py
+1
View File
@@ -19,6 +19,7 @@ class ConversionSession:
conversion_id: str
source_filename: str
resolution: conversion_pb2.ConversionResolution
bucket_name: str
upload_object_key: str
status: conversion_pb2.ConversionStatus
python/packages/server/src/officeconvert_server/service.py
+42 -3
View File
@@ -22,6 +22,11 @@ from officeconvert.conversion import (
PHASE_EXTRACTING_NOTES,
PHASE_PDF_TO_IMAGES,
PHASE_PPTX_TO_PDF,
RESOLUTION_FHD,
RESOLUTION_HD,
RESOLUTION_QHD,
RESOLUTION_SD,
RESOLUTION_UHD,
)
from officeconvertapi.v1 import conversion_connect, conversion_pb2
@@ -31,6 +36,14 @@ from officeconvert_server.storage import S3Store
logger = logging.getLogger("uvicorn.error")
_RESOLUTION_PRESET_BY_PROTO = {
conversion_pb2.CONVERSION_RESOLUTION_SD: RESOLUTION_SD,
conversion_pb2.CONVERSION_RESOLUTION_HD: RESOLUTION_HD,
conversion_pb2.CONVERSION_RESOLUTION_FHD: RESOLUTION_FHD,
conversion_pb2.CONVERSION_RESOLUTION_QHD: RESOLUTION_QHD,
conversion_pb2.CONVERSION_RESOLUTION_UHD: RESOLUTION_UHD,
}
class ConversionServiceImpl(conversion_connect.ConversionService):
"""Implements the conversion API with in-memory state and S3 orchestration."""
@@ -54,6 +67,11 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
raise ConnectError(Code.INVALID_ARGUMENT, "source_filename is required")
if not source_filename.lower().endswith(".pptx"):
raise ConnectError(Code.INVALID_ARGUMENT, "only .pptx input is supported")
resolution = request.resolution
if resolution == conversion_pb2.CONVERSION_RESOLUTION_UNSPECIFIED:
resolution = conversion_pb2.CONVERSION_RESOLUTION_FHD
if resolution not in _RESOLUTION_PRESET_BY_PROTO:
raise ConnectError(Code.INVALID_ARGUMENT, "resolution is invalid")
conversion_id = str(uuid.uuid4())
bucket_name = f"oc-{conversion_id}"
@@ -70,6 +88,7 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
session = ConversionSession(
conversion_id=conversion_id,
source_filename=source_filename,
resolution=resolution,
bucket_name=bucket_name,
upload_object_key=upload_key,
status=conversion_pb2.CONVERSION_STATUS_PENDING,
@@ -186,11 +205,11 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
"""Execute conversion flow and persist terminal state in memory."""
started_at = time.monotonic()
logger.info(
"Starting conversion conversion_id=%s source_filename=%s dpi=%d "
"Starting conversion conversion_id=%s source_filename=%s resolution=%s "
"timeout_caps_s[pptx_to_pdf_total=%d,pdf_to_images_total=%d]",
session.conversion_id,
session.source_filename,
self._config.conversion_image_dpi,
conversion_pb2.ConversionResolution.Name(session.resolution),
self._config.conversion_pptx_to_pdf_timeout_seconds,
self._config.conversion_pdf_to_images_timeout_seconds,
)
@@ -210,7 +229,7 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
convert_pptx_to_slidedeck,
source_path,
work_dir,
dpi=self._config.conversion_image_dpi,
resolution=_RESOLUTION_PRESET_BY_PROTO[session.resolution],
pptx_to_pdf_timeout_s=self._config.conversion_pptx_to_pdf_timeout_seconds,
pdf_to_images_timeout_s=self._config.conversion_pdf_to_images_timeout_seconds,
pptx_to_pdf_base_timeout_s=self._config.conversion_pptx_to_pdf_base_timeout_seconds,
@@ -224,6 +243,20 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
max_progress=max_value,
),
)
logger.info(
"Resolved conversion plan conversion_id=%s source_filename=%s "
"resolution=%s inferred_dpi=%d output_size=%dx%d "
"computed_timeouts_s[pptx_to_pdf_total=%d,pdf_to_images_total=%d,pdf_to_images_per_page=%d]",
session.conversion_id,
session.source_filename,
conversion_pb2.ConversionResolution.Name(session.resolution),
result.inferred_dpi,
result.width,
result.height,
result.pptx_to_pdf_timeout_s,
result.pdf_to_images_timeout_s,
result.pdf_to_images_page_timeout_s,
)
self._set_session_progress(
session,
phase=conversion_pb2.CONVERSION_PHASE_UPLOADING_RESULTS,
@@ -235,6 +268,8 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
session,
result.slides,
result.source_filename,
result.width,
result.height,
lambda current, max_value: self._set_session_progress(
session,
phase=conversion_pb2.CONVERSION_PHASE_UPLOADING_RESULTS,
@@ -300,6 +335,8 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
session: ConversionSession,
slides: list[SlideArtifact],
source_filename: str,
width: int,
height: int,
progress_callback: Callable[[int, int], None] | None = None,
) -> conversion_pb2.SlideDeck:
"""Upload generated slide images and construct API response payload."""
@@ -328,6 +365,8 @@ class ConversionServiceImpl(conversion_connect.ConversionService):
source_filename=source_filename,
slides=response_slides,
created_at=_to_timestamp(utc_now()),
width=width,
height=height,
)
async def _delayed_cleanup(self, session: ConversionSession) -> None: