Improve pretty-printing of results and fix other issues

Author: junyuan-chen

Project.toml

@@ -4,13 +4,11 @@ authors = ["Junyuan Chen <norman.chen.git@gmail.com>"]
 version = "0.1.0"
 
 [deps]
-Combinatorics = "861a8166-3701-5b0c-9a16-15d98fcdc6aa"
 Dictionaries = "85a47980-9c8c-11e8-2b9f-f7ca1fa99fb4"
 DiffinDiffsBase = "7fc23ce2-6e83-4e3c-822f-a79085ccb3e8"
 FixedEffectModels = "9d5cd8c9-2029-5cab-9928-427838db53e3"
 FixedEffects = "c8885935-8500-56a7-9867-7708b20db0eb"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
-Printf = "de0858da-6303-5e67-8744-51eddeeeb8d7"
 Reexport = "189a3867-3050-52da-a836-e630ba90ab69"
 SplitApplyCombine = "03a91e81-4c3e-53e1-a0a4-9c0c8f19dd66"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
@@ -21,7 +19,6 @@ TypedTables = "9d95f2ec-7b3d-5a63-8d20-e2491e220bb9"
 Vcov = "ec2bfdc2-55df-4fc9-b9ae-4958c2cf2486"
 
 [compat]
-Combinatorics = "1"
 Dictionaries = "0.3"
 DiffinDiffsBase = "0.2"
 FixedEffectModels = "1.4"

src/InteractionWeightedDIDs.jl

@@ -1,16 +1,14 @@
 module InteractionWeightedDIDs
 
 using Base: Callable
-using Combinatorics: combinations
 using Dictionaries: Dictionary
 using FixedEffectModels: FixedEffectTerm, Combination,
-    fe, has_fe, parse_fixedeffect, basecol, isnested, nunique
+    fe, parse_fixedeffect, basecol, isnested, nunique
 using FixedEffects
 using LinearAlgebra: Factorization, Symmetric, cholesky!
-using Printf
 using Reexport
 using SplitApplyCombine: group, mapview
-using StatsBase: AbstractWeights, CovarianceEstimator, UnitWeights, NoQuote, vcov
+using StatsBase: AbstractWeights, CovarianceEstimator, UnitWeights, PValue, TestStat, NoQuote
 using StatsFuns: fdistccdf
 using StatsModels: coefnames
 using Tables: getcolumn, columnnames
@@ -43,6 +41,5 @@ export CheckVcov,
 include("utils.jl")
 include("procedures.jl")
 include("did.jl")
-include("lsweights.jl")
 
 end

src/did.jl

@@ -66,67 +66,74 @@ function result(::Type{Reg}, @nospecialize(nt::NamedTuple))
     return merge(nt, (result=didresult,))
 end
 
+"""
+    has_fe(r::RegressionBasedDIDResult)
+
+Test whether any fixed effect is involved in regression.
+"""
 has_fe(r::RegressionBasedDIDResult) = !isempty(r.fenames)
 
-format_scientific(f::Number) = @sprintf("%.3f", f)
-
-function _summary(r::RegressionBasedDIDResult)
-    top = ["Number of obs" sprint(show, nobs(r), context=:compact=>true);
-        "Degrees of freedom" sprint(show, nobs(r) - dof_residual(r), context=:compact=>true);
-        "F-statistic" sprint(show, r.F, context = :compact => true);
-        "p-value" format_scientific(r.p)]
-    fe_info = has_fe(r) ?  
-        ["Converged" sprint(show, r.feconverged, context=:compact=>true);
-        "Singletons dropped" sprint(show, r.nfesingledropped, context=:compact=>true);
-        ] : nothing
-    return top, fe_info
-end
+_top_info(r::RegressionBasedDIDResult) = (
+    "Number of obs" => nobs(r),
+    "Degrees of freedom" => nobs(r) - dof_residual(r),
+    "F-statistic" => TestStat(r.F),
+    "p-value" => PValue(r.p)
+)
 
 _nunique(t::Table, s::Symbol) = length(unique(getproperty(t, s)))
 
-_excluded_rel_str(tr::DynamicTreatment) = 
-    isempty(tr.exc) ? "None" : join(string.(tr.exc), " ")
+_excluded_rel_str(tr::DynamicTreatment) =
+    isempty(tr.exc) ? "none" : join(string.(tr.exc), " ")
+
+_treat_info(r::RegressionBasedDIDResult{true}, tr::DynamicTreatment) = (
+    "Number of cohorts" => _nunique(r.treatinds, r.treatname),
+    "Interactions within cohorts" => length(columnnames(r.treatinds)) - 2,
+    "Relative time periods" => _nunique(r.treatinds, :rel),
+    "Excluded periods" => NoQuote(_excluded_rel_str(tr))
+)
 
-_treat_info(tr::DynamicTreatment, trinds::Table, treatname::Symbol) =
-    ["Number of cohorts" sprint(show, _nunique(trinds, treatname), context=:compact=>true);
-    "Interactions within cohorts" sprint(show, length(columnnames(trinds))-2, context=:compact=>true);
-    "Relative time periods" sprint(show, _nunique(trinds, :rel), context=:compact=>true);
-    "Excluded periods" sprint(show, NoQuote(_excluded_rel_str(tr)), context=:compact=>true)]
+_treat_info(r::RegressionBasedDIDResult{false}, tr::DynamicTreatment) = (
+    "Relative time periods" => _nunique(r.treatinds, :rel),
+    "Excluded periods" => NoQuote(_excluded_rel_str(tr))
+)
 
 _treat_spec(r::RegressionBasedDIDResult{true}, tr::DynamicTreatment{SharpDesign}) =
     "Cohort-interacted sharp dynamic specification"
 
 _treat_spec(r::RegressionBasedDIDResult{false}, tr::DynamicTreatment{SharpDesign}) =
     "Sharp dynamic specification"
-    
-function show(io::IO, r::RegressionBasedDIDResult;
-    totalwidth::Int=70, interwidth::Int=4+mod(totalwidth,2))
+
+_fe_info(r::RegressionBasedDIDResult) = (
+    "Converged" => r.feconverged,
+    "Singletons dropped" => r.nfesingledropped
+)
+
+show(io::IO, ::RegressionBasedDIDResult) = print(io, "Regression-based DID result")
+
+function show(io::IO, ::MIME"text/plain", r::RegressionBasedDIDResult;
+        totalwidth::Int=70, interwidth::Int=4+mod(totalwidth,2))
     halfwidth = div(totalwidth-interwidth, 2)
-    top, fe_info = _summary(r)
-    tr_info = _treat_info(r.tr, r.treatinds, r.treatname)
-    blocks = [top, tr_info, fe_info]
+    top_info = _top_info(r)
+    fe_info = has_fe(r) ? _fe_info(r) : ()
+    tr_info = _treat_info(r, r.tr)
+    blocks = (top_info, tr_info, fe_info)
     fes = has_fe(r) ? join(string.(r.fenames), " ") : "none"
     fetitle = string("Fixed effects: ", fes)
-    blocktitles = ["Summary of results",
-                   _treat_spec(r, r.tr),
-                   fetitle[1:min(totalwidth,length(fetitle))]]
+    blocktitles = ("Summary of results: Regression-based DID",
+        _treat_spec(r, r.tr), fetitle[1:min(totalwidth,length(fetitle))])
+
     for (ib, b) in enumerate(blocks)
-        for i in 1:size(b, 1)
-            b[i, 1] = b[i, 1] * ":"
-        end
-        println(io, "─" ^totalwidth)
+        println(io, repeat('─', totalwidth))
         println(io, blocktitles[ib])
-        println(io, "─" ^totalwidth)
-        for i in 1:(div(size(b, 1) - 1, 2)+1)
-            print(io, b[2*i-1, 1])
-            print(io, lpad(b[2*i-1, 2], halfwidth - length(b[2*i-1, 1])))
-            print(io, " " ^interwidth)
-            if size(b, 1) >= 2*i
-                print(io, b[2*i, 1])
-                print(io, lpad(b[2*i, 2], halfwidth - length(b[2*i, 1])))
+        if !isempty(b)
+            # Print line between block title and block content
+            println(io, repeat('─', totalwidth))
+            for (i, e) in enumerate(b)
+                print(io, e[1], ':')
+                print(io, lpad(e[2], halfwidth - length(e[1]) - 1))
+                print(io, isodd(i) ? repeat(' ', interwidth) : '\n')
             end
-            println(io)
         end
     end
-    println(io, "─" ^totalwidth)
+    print(io, repeat('─', totalwidth))
 end

src/procedures.jl

@@ -188,14 +188,16 @@ end
 function _genindicator(idx::Vector{Int}, esample::BitVector, n::Int)
     v = zeros(n)
     iv, r = 1, 1
+    nr = length(idx)
     @inbounds for i in eachindex(esample)
         if esample[i]
-            if i == idx[r]
+            if idx[r] == i
                 v[iv] = 1.0
                 r += 1
             end
             iv += 1
         end
+        r > nr && break
     end
     return v
 end
@@ -378,16 +380,15 @@ function estvcov(data, esample::BitVector, vce::CovarianceEstimator, coef::Vecto
     has_intercept = has_intercept || has_fe_intercept
     df_F = max(1, Vcov.df_FStat(vce_data, concrete_vce, has_intercept))
     p = fdistccdf(max(length(coef) - has_intercept, 1), df_F, F)
-
-    return (vcov_mat=vcov_mat, dof_resid=dof_resid, F=F, p=p), true
+    return (vcov_mat=vcov_mat, vce=concrete_vce, dof_resid=dof_resid, F=F, p=p), true
 end
 
 """
     EstVcov <: StatsStep
 
 Call [`InteractionWeightedDIDs.estvcov`](@ref) to
 estimate variance-covariance matrix and F-statistic.
-The returned objects named `vcov_mat`, `dof_resid`, `F` and `p`
+The returned objects named `vcov_mat`, `vce`, `dof_resid`, `F` and `p`
 may be shared across multiple specifications.
 """
 const EstVcov = StatsStep{:EstVcov, typeof(estvcov)}

test/did.jl

@@ -1,8 +1,8 @@
 @testset "RegressionBasedDID" begin
     hrs = exampledata("hrs")
-    r = @did(Reg, data=hrs, dynamic(:wave, -1), notyettreated([11]), vce=Vcov.cluster(:hhidpn),
-        yterm=term(:oop_spend), treatname=:wave_hosp, treatintterms=(),
-        xterms=(fe(:wave)+fe(:hhidpn)))
+    r = @did(Reg, data=hrs, dynamic(:wave, -1), notyettreated([11]),
+        vce=Vcov.cluster(:hhidpn), yterm=term(:oop_spend), treatname=:wave_hosp,
+        treatintterms=(), xterms=(fe(:wave)+fe(:hhidpn)))
     @test coef(r, "rel: -3 & wave_hosp: 10") ≈ 591.04639 atol=1e-5
     @test coef(r, "rel: -2 & wave_hosp: 9") ≈ 298.97735 atol=1e-5
     @test coef(r, "rel: -2 & wave_hosp: 10") ≈ 410.58102 atol=1e-5
@@ -14,4 +14,38 @@
     @test coef(r, "rel: 2 & wave_hosp: 8") ≈ 800.10647 atol=1e-5
     @test nobs(r) == 2624
 
+    @test sprint(show, r) == "Regression-based DID result"
+    @test sprint(show, MIME("text/plain"), r) == """
+        ──────────────────────────────────────────────────────────────────────
+        Summary of results: Regression-based DID
+        ──────────────────────────────────────────────────────────────────────
+        Number of obs:               2624    Degrees of freedom:           670
+        F-statistic:                 4.81    p-value:                    <1e-5
+        ──────────────────────────────────────────────────────────────────────
+        Cohort-interacted sharp dynamic specification
+        ──────────────────────────────────────────────────────────────────────
+        Number of cohorts:              3    Interactions within cohorts:    0
+        Relative time periods:          5    Excluded periods:              -1
+        ──────────────────────────────────────────────────────────────────────
+        Fixed effects: fe_wave fe_hhidpn
+        ──────────────────────────────────────────────────────────────────────
+        Converged:                   true    Singletons dropped:             0
+        ──────────────────────────────────────────────────────────────────────"""
+
+    r = @did(Reg, data=hrs, dynamic(:wave, -1), notyettreated([11]),
+        vce=Vcov.cluster(:hhidpn), yterm=term(:oop_spend), treatname=:wave_hosp,
+        treatintterms=(), cohortinteracted=false)
+    @test sprint(show, MIME("text/plain"), r) == """
+        ──────────────────────────────────────────────────────────────────────
+        Summary of results: Regression-based DID
+        ──────────────────────────────────────────────────────────────────────
+        Number of obs:               2624    Degrees of freedom:             6
+        F-statistic:                12.50    p-value:                   <1e-10
+        ──────────────────────────────────────────────────────────────────────
+        Sharp dynamic specification
+        ──────────────────────────────────────────────────────────────────────
+        Relative time periods:          5    Excluded periods:              -1
+        ──────────────────────────────────────────────────────────────────────
+        Fixed effects: none
+        ──────────────────────────────────────────────────────────────────────"""
 end

test/runtests.jl

@@ -17,6 +17,9 @@ import Base: ==
 ==(x::FixedEffect{R,I}, y::FixedEffect{R,I}) where {R,I} =
     x.refs == y.refs && x.interaction == y.interaction && x.n == y.n
 
+# A workaround to be replaced
+==(x::Vcov.ClusterCovariance, y::Vcov.ClusterCovariance) = true
+
 const tests = [
     "procedures",
     "did"