Skip to content

Commit

Permalink
adding notebooks, data, etc
Browse files Browse the repository at this point in the history
  • Loading branch information
@spiralizing
spiralizing committed Aug 6, 2023
1 parent 2fb4030 commit b951cff
Show file tree
Hide file tree
Showing 144 changed files with 3,907,454 additions and 25,567 deletions.
269 changes: 18 additions & 251 deletions BeethovenQuartets.ipynb
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@
"cells": [
{
"cell_type": "code",
"execution_count": 2,
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
Expand All @@ -21,7 +21,7 @@
},
{
"cell_type": "code",
"execution_count": 3,
"execution_count": 2,
"metadata": {},
"outputs": [
{
Expand All @@ -35,12 +35,12 @@
}
],
"source": [
"m21 = pyimport(\"music21\") #music21 manages to load successfully the musicxml files, while musicxml Julia package doesn't."
"m21 = pyimport(\"music21\") "
]
},
{
"cell_type": "code",
"execution_count": 4,
"execution_count": 3,
"metadata": {},
"outputs": [
{
Expand All @@ -55,7 +55,7 @@
],
"source": [
"#functions from personal packages.\n",
"git_path = \"/home/alfredo/Git/CoE-Testing/src/\"\n",
"git_path = \"/home/alfredo/Git/Harmony_Evolution/src\"\n",
"include(joinpath(git_path,\"CEGFunctions.jl\"))\n",
"include(joinpath(git_path,\"Constr_series.jl\"))\n",
"include(joinpath(git_path,\"IM_Functions.jl\"))\n",
Expand All @@ -64,264 +64,31 @@
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"get_key_ent (generic function with 1 method)"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"#More functions...\n",
"function get_xml_df(piece_xml)\n",
" piece = get_piece_by_measure(piece_xml, csv=false)\n",
" df_piece = DataFrame(\n",
" :Measure => convert(Array{Int64,1},piece[:,1]),\n",
" :TimeSignature => piece[:,2],\n",
" :StartQuarter => piece[:,3],\n",
" :EndQuarter => piece[:,4],\n",
" :Duration => piece[:,5],\n",
" :Pitch => convert(Array{Int64,1},piece[:,6])\n",
" )\n",
" return df_piece\n",
"end\n",
"function get_csv_df(piece_csv)\n",
" piece = get_piece_by_measure(piece_csv, csv=true)[1]\n",
" num_mea = []\n",
" for nm in 1:length(piece)\n",
" push!(num_mea,[nm for i in 1:size(piece[nm],1)])\n",
" end\n",
" num_mea = vcat(num_mea...)\n",
" piece = vcat(piece...)\n",
" df_piece = DataFrame(\n",
" :Measure => convert(Array{Int64,1},num_mea),\n",
" :TimeSignature => piece[:,2],\n",
" :StartTime => piece[:,3],\n",
" :EndTime => piece[:,4],\n",
" :Duration => piece[:,5],\n",
" :Pitch => convert(Array{Int64,1},piece[:,6])\n",
" )\n",
" return df_piece\n",
"end\n",
"function fastuniq(v)\n",
" v1 = Vector{eltype(v)}()\n",
" if length(v)>0\n",
" laste = v[1]\n",
" push!(v1,laste)\n",
" for e in v\n",
" if e != laste\n",
" laste = e\n",
" push!(v1,laste)\n",
" end\n",
" end\n",
" end\n",
" return v1\n",
"end\n",
"get_entropy(probs) = mapreduce(x-> - x * log2(x), +, probs)\n",
"function get_key_IC(list_keys)\n",
" prob_keys = convert(Array{Float64,1}, list_keys[:,2])\n",
" return - log2(prob_keys[1])\n",
"end\n",
"function get_distance_ces(ce1, ce2)\n",
" z_dif = ce2[3] - ce1[3]\n",
" while abs(z_dif) > h_octav / 2 #translating over z to be in the same octave (same CE region)\n",
" if z_dif > 0\n",
" ce2[3] = ce2[3] - h_octav\n",
" else\n",
" ce2[3] = ce2[3] + h_octav\n",
" end\n",
" z_dif = ce2[3] - ce1[3]\n",
" end\n",
" return round(euclidean(ce1,ce2), digits = 4)\n",
"end\n",
"function divide_chunk_notes(chunk, w_s, n_c)\n",
" st_m = minimum(chunk[:,3])\n",
" c_out = []\n",
" for i in 1:n_c\n",
" push!(c_out, chunk[findall(x -> st_m + w_s * (i - 1) <= x < st_m + w_s * i, chunk[:,3]),:])\n",
" end\n",
" return filter(x-> !isempty(x),c_out)\n",
"end\n",
"function get_distance_to_keys(c_i)\n",
" d_to_keys = round.(map(x-> euclidean(c_i, x), pos_all_keys), digits = 4) #computing the eclidean distance to all keys\n",
"\n",
" ranking = sortperm(d_to_keys) #ranking the distances from the closest to the farthest\n",
"\n",
" return [all_keys[ranking] d_to_keys[ranking] pos_all_keys[ranking]][1:12,:]\n",
"end\n",
"function cluster_notes(ptcs)\n",
" p_m12 = map(x-> mod(x,12),ptcs)\n",
" spi_notes = get_cfpitch(p_m12)\n",
" low_notes = findall(x-> x<6, spi_notes)\n",
" high_notes = findall(x-> x>=6, spi_notes)\n",
" if !isempty(high_notes) && !isempty(low_notes)\n",
" if length(high_notes) < length(low_notes)\n",
" for i in 1:length(high_notes)\n",
" spi_notes = shift_outlier(spi_notes, high_notes[i])\n",
" end\n",
" else\n",
" for i in 1:length(low_notes)\n",
" spi_notes = shift_outlier(spi_notes, low_notes[i])\n",
" end\n",
" end\n",
" end\n",
" dmean = Float64[]\n",
" oliers = Int64[]\n",
" new_spi = []\n",
" conv = false\n",
" while conv == false\n",
" dt_mean = zeros(length(spi_notes))\n",
" for i in 1:length(spi_notes)\n",
" dt_mean[i] = abs(spi_notes[i] - mean(spi_notes[1:end .!= i]))\n",
" end\n",
" d, olier = findmax(dt_mean)\n",
" push!(dmean, mean(dt_mean)); push!(oliers, olier); push!(new_spi, spi_notes)\n",
" #println(mean(dt_mean),'\\t', olier, '\\t', p_cf)\n",
" #println(mean(dt_mean),'\\t', olier)\n",
" if length(oliers) > 10 && length(unique(oliers[end-4:end])) <= 2\n",
" conv = true\n",
" break\n",
" end\n",
" spi_notes = shift_outlier(spi_notes, olier)\n",
" end\n",
" return new_spi[findmin(dmean)[2]]\n",
"end\n",
"function shift_outlier(notes, olier)\n",
" dif = notes[olier] - median(notes)\n",
" notes_new = copy(notes)\n",
" if dif > 0\n",
" notes_new[olier] = notes_new[olier] - 12\n",
" elseif dif < 0\n",
" notes_new[olier] = notes_new[olier] + 12\n",
" end\n",
" return notes_new\n",
"end\n",
"function get_center_effect(chunk_notes; r=1, h=sqrt(2/15), mod_12=false,all_keys=all_keys, pos_all_keys=pos_all_keys, sbeat_w=[[1.],[1.]], lin_w=1) \n",
" ptcs = chunk_notes[:,6]\n",
" durs = chunk_notes[:,5]\n",
" pbeat = chunk_notes[:,1]\n",
" beat_w = ones(length(durs)) #array of the beat weights\n",
" for b = 1:length(sbeat_w[1])\n",
" loc_b = findall(x-> x==sbeat_w[1][b], pbeat) #finding all notes that start at beat sbeat_w[1][b]\n",
" beat_w[loc_b] .= sbeat_w[2][b] #this is the weight.\n",
" end\n",
" notas, n_we = get_local_lin_w(ptcs, lin_w) #doing the linear weight in the pitches\n",
" ii = vcat(map(x-> findall(y-> y==x, notas), ptcs)...)\n",
" #println(ptcs)\n",
" b_wei = n_we[ii] #getting the linear weight for every note i n the array of pitches\n",
" ###DO A FUNCTION FROM HERE\n",
" spi_ix = cluster_notes(ptcs) .+ 24\n",
" #TO HERE\n",
" spi_p = map(x-> get_pitch(x, r=r, h=h), spi_ix) #getting the location (x,y,z) for each pitch\n",
" t_ws = map((x,y,z)-> x*y*z, beat_w,durs, b_wei) #computing the total weights\n",
" cv_i = map((x,y)-> x*y, t_ws, spi_p) / sum(t_ws) #computing the location of the pitches with their relative weights\n",
" c_i = sum(cv_i) #finding the center of effect\n",
" return c_i\n",
"end\n",
"function get_key_ent(key_list)\n",
" dists = map(x-> exp(-12 * x), Float64.(key_list[:,2]))\n",
" key_p = dists ./ sum(dists)\n",
" return get_entropy(key_p)\n",
"end"
]
},
{
"cell_type": "code",
"execution_count": 6,
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"xml_path = \"/home/alfredo/MusicPenn/Beethoven_HAnalysis/XMLFiles\" #xml files \n",
"adata_path = \"/home/alfredo/MusicPenn/Beethoven_HAnalysis/AnnotationsTSV\" #annotated data\n",
"out_path = \"/home/alfredo/MusicPenn/Beethoven_HAnalysis/KeyCallCSV\"\n",
"adata_path = \"/home/alfredo/Git/Harmony_Evolution/Beethoven_Annotations/\" #annotated data\n",
"out_path = \"/home/alfredo/Git/Harmony_Evolution/Beethoven_Annotations/\"\n",
"xml_list = readdir(xml_path)\n",
"adata_list = readdir(adata_path);"
"ann_files = readdir(adata_path)\n",
"adata_list = ann_files[findall(x -> occursin(r\".tsv\", x), ann_files)];"
]
},
{
"cell_type": "code",
"execution_count": 7,
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"70-element Vector{String}:\n",
" \"op127_no12\"\n",
" \"op127_no12\"\n",
" \"op127_no12\"\n",
" \"op127_no12\"\n",
" \"op130_no13\"\n",
" \"op130_no13\"\n",
" \"op130_no13\"\n",
" \"op130_no13\"\n",
" \"op130_no13\"\n",
" \"op130_no13\"\n",
" ⋮\n",
" \"op59_no9\"\n",
" \"op74_no10\"\n",
" \"op74_no10\"\n",
" \"op74_no10\"\n",
" \"op74_no10\"\n",
" \"op95_no11\"\n",
" \"op95_no11\"\n",
" \"op95_no11\"\n",
" \"op95_no11\""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"outputs": [],
"source": [
"opus_number = map(x-> join(split(x, \"_\")[1:2],\"_\"), xml_list)\n",
"u_opus = unique(opus_number);"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"16-element Vector{Vector{Int64}}:\n",
" [1, 2, 3, 4]\n",
" [5, 6, 7, 8, 9, 10]\n",
" [11, 12, 13, 14, 15, 16, 17]\n",
" [18, 19, 20, 21, 22]\n",
" [23, 24, 25, 26]\n",
" [27, 28, 29, 30]\n",
" [31, 32, 33, 34]\n",
" [35, 36, 37, 38]\n",
" [39, 40, 41, 42]\n",
" [43, 44, 45, 46]\n",
" [47, 48, 49, 50]\n",
" [51, 52, 53, 54]\n",
" [55, 56, 57, 58]\n",
" [59, 60, 61, 62]\n",
" [63, 64, 65, 66]\n",
" [67, 68, 69, 70]"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"ix_all = [findall(x-> occursin(u_opus[i], x), adata_list) for i in 1:length(u_opus)]"
]
},
{
"cell_type": "code",
"execution_count": 50,
"execution_count": 6,
"metadata": {},
"outputs": [
{
Expand Down Expand Up @@ -816,7 +583,6 @@
}
],
"source": [
"\n",
"for xml_i in 1:length(xml_list)\n",
" local fh_kseq\n",
" #loading csv from midi\n",
Expand Down Expand Up @@ -889,12 +655,13 @@
" name_out = \"CoEKeyCallNEW-$(join([split(xml_file,\".\")[1] \"csv\"],\".\"))\"\n",
" CSV.write(joinpath(out_path,name_out), df_out, header=true)\n",
" println(\"File $(xml_file) DONE!\")\n",
" flush(stdout)\n",
"end"
]
},
{
"cell_type": "code",
"execution_count": 51,
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
Expand All @@ -904,7 +671,7 @@
},
{
"cell_type": "code",
"execution_count": 52,
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
Expand Down Expand Up @@ -943,14 +710,14 @@
},
{
"cell_type": "code",
"execution_count": 64,
"execution_count": 11,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"the median accuracy for all the movements is: 67.74%\n"
"the median accuracy for all the movements is: 68.435%\n"
]
}
],
Expand Down
Loading

0 comments on commit b951cff

Please sign in to comment.