Enhance Blob handling and MIME type integration, add solvable to summary level and rm Blobentries (#1213)

* add solvable to summary level, rm blobentries

* Enhance Blob handling and MIME type integration with new functions and tests

* fix docs

* Apply suggestions from code review

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* fixes and formatting

---------

Co-authored-by: Johannes Terblanche <Affie@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Author: Affie

Project.toml

@@ -14,6 +14,7 @@ Graphs = "86223c79-3864-5bf0-83f7-82e725a168b6"
 InteractiveUtils = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
 JSON = "682c06a0-de6a-54ab-a142-c8b1cf79cde6"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
+MIMEs = "6c6e2e6c-3030-632d-7369-2d6c69616d65"
 ManifoldsBase = "3362f125-f0bb-47a3-aa74-596ffd7ef2fb"
 OrderedCollections = "bac558e1-5e72-5ebc-8fee-abe8a469f55d"
 Pkg = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
@@ -55,6 +56,7 @@ InteractiveUtils = "1.11"
 JSON = "1.0.0"
 LieGroups = "0.1"
 LinearAlgebra = "1.11"
+MIMEs = "1.1"
 ManifoldsBase = "1, 2"
 OrderedCollections = "1.4"
 Pkg = "1.4, 1.5"

ext/BlobArrow.jl

@@ -2,19 +2,20 @@ module BlobArrow
 
 using Arrow
 using DistributedFactorGraphs
-using DistributedFactorGraphs: _MIMETypes
+using DistributedFactorGraphs: _MIMEOverrides, getMimetype
 
 function __init__()
     @info "Including Arrow blobs support in DFG."
-    return push!(_MIMETypes, MIME("application/vnd.apache.arrow.file") => format"Arrow") # see issue #507
+    push!(_MIMEOverrides, format"Arrow" => MIME("application/vnd.apache.arrow.file"))
+    return nothing
 end
 
 # kwargs: compress = :lz4,
 function DFG.packBlob(::Type{format"Arrow"}, data; kwargs...)
     io = IOBuffer()
     Arrow.write(io, data; kwargs...)
     blob = take!(io)
-    mimetype = findfirst(==(format"Arrow"), _MIMETypes)
+    mimetype = getMimetype(format"Arrow")
     return blob, mimetype
 end
 

src/Common.jl

@@ -214,7 +214,7 @@ function Timestamp(epoch::Val{:unix}, t::Float64, zone = tz"UTC")
 end
 Timestamp(t::Float64, zone = tz"UTC") = Timestamp(Val(:unix), t, zone)
 function Timestamp(epoch::Val{:rata}, t::Float64, zone = tz"UTC")
-    return TimeDateZone(convert(DateTime, Millisecond(t*10^3)), zone)
+    return TimeDateZone(convert(DateTime, Millisecond(t * 10^3)), zone)
 end
 
 function now_tdz(zone = tz"UTC")

src/DataBlobs/entities/BlobEntry.jl

@@ -20,23 +20,23 @@ StructUtils.@kwarg struct Blobentry
     """ (Optional) crc32c hash value to ensure data consistency which must correspond to the stored hash upon retrieval."""
     crchash::Union{UInt32, Nothing} =
         nothing & (
-            json=(
-                lower = h->isnothing(h) ? nothing : string(h, base = 16),
-                lift = s->isnothing(s) ? nothing : parse(UInt32, s; base = 16),
+            json = (
+                lower = h -> isnothing(h) ? nothing : string(h; base = 16),
+                lift = s -> isnothing(s) ? nothing : parse(UInt32, s; base = 16),
             )
         )
     """ (Optional) sha256 hash value to ensure data consistency which must correspond to the stored hash upon retrieval."""
     shahash::Union{Vector{UInt8}, Nothing} =
         nothing & (
-            json=(
-                lower = h->isnothing(h) ? nothing : bytes2hex(h),
-                lift = s->isnothing(s) ? nothing : hex2bytes(s),
+            json = (
+                lower = h -> isnothing(h) ? nothing : bytes2hex(h),
+                lift = s -> isnothing(s) ? nothing : hex2bytes(s),
             )
         )
     """ Source system or application where the blob was created (e.g., webapp, sdk, robot)"""
     origin::String = ""
     """Number of bytes in blob serialized as a string"""
-    size::Int64 = -1 & (json=(lower = string, lift = x->parse(Int64, x)))
+    size::Int64 = -1 & (json = (lower = string, lift = x -> parse(Int64, x)))
     """ Additional information that can help a different user of the Blob. """
     description::String = ""
     """ MIME description describing the format of binary data in the `Blob`, e.g. 'image/png' or 'application/json'. """

src/DataBlobs/services/BlobPacking.jl

@@ -1,49 +1,91 @@
-# using FileIO
-# using ImageIO
-# using LasIO
-# using BSON
-# using OrderedCollections
-
-# 2 types for now with MIME type
-# 1. JSON     - application/octet-stream/json
-# 2. FileIO   - application/octet-stream 
-#             - application/bson 
-#             - image/jpeg
-#             - image/png
-#             - application/vnd.apache.arrow.file
-
-const _MIMETypes = OrderedDict{MIME, DataType}()
-push!(_MIMETypes, MIME("application/octet-stream/json") => format"JSON")
-push!(_MIMETypes, MIME("application/bson") => format"BSON")
-push!(_MIMETypes, MIME("image/png") => format"PNG")
-push!(_MIMETypes, MIME("image/jpeg") => format"JPG")
-push!(_MIMETypes, MIME("application/vnd.las") => format"LAS")
-push!(_MIMETypes, MIME("application/vnd.apache.parque") => format"Parquet") # Provided by FileIO with ParquetFiles 
+##==============================================================================
+## BlobPacking: format <-> MIME type bridging and blob serialization
+##==============================================================================
+
+# Override dictionary for formats not covered by MIMEs.jl + FileIO auto-detection.
+# Standard types (PNG, JPEG, CSV, etc.) are auto-detected and don't need entries here.
+const _MIMEOverrides = OrderedDict{DataType, MIME}(
+    format"JSON" => MIME("application/json"),
+    format"BSON" => MIME("application/bson"),
+    format"LAS" => MIME("application/vnd.las"),
+    format"Parquet" => MIME("application/vnd.apache.parquet"),
+)
+
+"""
+    getMimetype(::Type{DataFormat{S}}) -> MIME
+
+Get the MIME type for a FileIO `DataFormat`. Uses FileIO's extension registry
+and MIMEs.jl for standard types, falls back to `_MIMEOverrides` for
+domain-specific formats.
+
+# Examples
+```julia
+getMimetype(format"PNG")  # MIME("image/png")
+getMimetype(format"JSON") # MIME("application/json")
+```
+"""
+function getMimetype(::Type{DataFormat{S}}) where {S}
+    T = DataFormat{S}
+    haskey(_MIMEOverrides, T) && return _MIMEOverrides[T]
+    try
+        finfo = FileIO.info(T)
+        ext = finfo[2]
+        ext = ext isa AbstractVector ? first(ext) : ext
+        m = mime_from_extension(ext)
+        !isnothing(m) && return m
+    catch
+    end
+    return MIME("application/octet-stream")
+end
+
+"""
+    getDataFormat(::MIME) -> Union{Type{DataFormat{S}}, Nothing}
+
+Get the FileIO `DataFormat` for a MIME type. Uses MIMEs.jl and FileIO's extension
+registry, falls back to `_MIMEOverrides`.
+
+Returns `nothing` if no matching format is found.
+
+# Examples
+```julia
+getDataFormat(MIME("image/png"))        # format"PNG"
+getDataFormat(MIME("application/json")) # format"JSON"
+```
+"""
+function getDataFormat(m::MIME)
+    for (fmt, mime) in _MIMEOverrides
+        mime == m && return fmt
+    end
+    ext = extension_from_mime(m)
+    sym = get(FileIO.ext2sym, ext, nothing)
+    !isnothing(sym) && return DataFormat{sym}
+    return nothing
+end
 
 """
     packBlob
-Convert a file (JSON, JPG, PNG, BSON, LAS) to Vector{UInt8} for use as a Blob.  
-Returns the blob and MIME type.
+Convert data to `Vector{UInt8}` for use as a Blob. Returns `(blob, mimetype)`.
+The MIME type is automatically determined from the DataFormat.
 """
 function packBlob end
+
 """
     unpackBlob
-Convert a Blob back to the origanal typ using the MIME type or DataFormat type.
+Convert a Blob back to the original type using the MIME type or DataFormat type.
 """
 function unpackBlob end
 
 unpackBlob(mime::String, blob) = unpackBlob(MIME(mime), blob)
 
 function unpackBlob(T::MIME, blob)
-    dataformat = get(_MIMETypes, T, nothing)
+    dataformat = getDataFormat(T)
     isnothing(dataformat) && error("Format not found for MIME type $(T)")
     return unpackBlob(dataformat, blob)
 end
 
 # 1. JSON strings are saved as is
 function packBlob(::Type{format"JSON"}, json_str::String)
-    mimetype = findfirst(==(format"JSON"), _MIMETypes)
-    # blob = codeunits(json_str)
+    mimetype = getMimetype(format"JSON")
     blob = Vector{UInt8}(json_str)
     return blob, mimetype
 end
@@ -55,16 +97,12 @@ end
 unpackBlob(entry::Blobentry, blob::Vector{UInt8}) = unpackBlob(entry.mimetype, blob)
 unpackBlob(eb::Pair{<:Blobentry, Vector{UInt8}}) = unpackBlob(eb[1], eb[2])
 
-# 2/ FileIO
+# 2. FileIO formats (PNG, JPEG, BSON, LAS, Parquet, etc.)
 function packBlob(::Type{T}, data::Any; kwargs...) where {T <: DataFormat}
     io = IOBuffer()
     save(Stream{T}(io), data; kwargs...)
     blob = take!(io)
-    mimetype = findfirst(==(T), _MIMETypes)
-    if isnothing(mimetype)
-        @warn "No MIME type found for format $T"
-        mimetype = MIME"application/octet-stream"
-    end
+    mimetype = getMimetype(T)
     return blob, mimetype
 end
 
@@ -73,18 +111,13 @@ function unpackBlob(::Type{T}, blob::Vector{UInt8}) where {T <: DataFormat}
     return load(Stream{T}(io))
 end
 
-# if false
-# json_str = "{\"name\":\"John\"}"
-# blob, mimetype = packBlob(format"JSON", json_str)
-# @assert json_str == unpackBlob(format"JSON", blob)
-# @assert json_str == unpackBlob(MIME("application/octet-stream/json"), blob)
-# @assert json_str == unpackBlob("application/octet-stream/json", blob)
-
-# blob,mime = packBlob(format"PNG", img)
-# up_img = unpackBlob(format"PNG", blob)
-
-# #TODO BSON does not work yet, can extend [un]packBlob(::Type{format"BSON"}, ...)
-# packBlob(format"BSON", Dict("name"=>"John"))
-# unpackBlob(format"BSON", Dict("name"=>"John"))
+"""    
+    getMimetype(io::IO) -> MIME
 
-# end
+Detect the MIME type of data in an IO stream using FileIO's format detection.
+"""
+function getMimetype(io::IO)
+    _getFormat(s::FileIO.Stream{T}) where {T} = T
+    stream = FileIO.query(io)
+    return getMimetype(_getFormat(stream))
+end

src/DataBlobs/services/BlobWrappers.jl

@@ -63,6 +63,27 @@ $(METHODLIST)
 """
 function deleteBlob_Agent! end
 
+"""
+Convenience wrapper to load a Blob for a given factor and Blobentry label.
+
+$(METHODLIST)
+"""
+function loadBlob_Factor end
+
+"""
+Convenience wrapper to save a Blob to a Blobstore and a Blobentry to a factor.
+
+$(METHODLIST)
+"""
+function saveBlob_Factor! end
+
+"""
+Convenience wrapper to delete a Blob from a Blobstore and its Blobentry from a factor.
+
+$(METHODLIST)
+"""
+function deleteBlob_Factor! end
+
 function loadBlob_Variable(
     dfg::AbstractDFG,
     variable_label::Symbol,
@@ -166,6 +187,42 @@ function deleteBlob_Agent!(dfg::AbstractDFG, entry_label::Symbol)
     return 2
 end
 
+function loadBlob_Factor(dfg::AbstractDFG, factor_label::Symbol, entry_label::Symbol)
+    entry = getFactorBlobentry(dfg, factor_label, entry_label)
+    blob = getBlob(dfg, entry)
+    return entry, blob
+end
+
+function saveBlob_Factor!(
+    dfg::AbstractDFG,
+    factor_label::Symbol,
+    blob::Vector{UInt8},
+    entry::Blobentry,
+)
+    addFactorBlobentry!(dfg, factor_label, entry)
+    addBlob!(dfg, entry, blob)
+    return entry
+end
+
+function saveBlob_Factor!(
+    dfg::AbstractDFG,
+    factor_label::Symbol,
+    blob::Vector{UInt8},
+    entry_label::Symbol,
+    blobstore::Symbol = :default;
+    blobentry_kwargs...,
+)
+    entry = Blobentry(entry_label, blobstore; blobentry_kwargs...)
+    return saveBlob_Factor!(dfg, factor_label, blob, entry)
+end
+
+function deleteBlob_Factor!(dfg::AbstractDFG, factor_label::Symbol, entry_label::Symbol)
+    entry = getFactorBlobentry(dfg, factor_label, entry_label)
+    deleteFactorBlobentry!(dfg, factor_label, entry_label)
+    deleteBlob!(dfg, entry)
+    return 2
+end
+
 function saveImage_Variable!(
     dfg::AbstractDFG,
     variable_label::Symbol,
@@ -174,19 +231,19 @@ function saveImage_Variable!(
     blobstore::Symbol = :default;
     entry_kwargs...,
 )
-    mimeType = get(entry_kwargs, :mimeType, MIME("image/png"))
-    format = _MIMETypes[mimeType]
-
-    blob, mimeType = packBlob(format, img)
+    mimetype = get(entry_kwargs, :mimeType, MIME("image/png"))
+    format = getDataFormat(mimetype)
+    isnothing(format) &&
+        throw(ArgumentError("Unsupported MIME type for image: $(mimetype)"))
+    blob, mimetype = packBlob(format, img)
 
-    size = string(length(blob))
     entry = Blobentry(
         entry_label,
         blobstore;
         blobid = uuid4(),
         entry_kwargs...,
-        size,
-        mimeType = string(mimeType),
+        size = length(blob),
+        mimetype,
     )
 
     return saveBlob_Variable!(dfg, variable_label, blob, entry)