From 6d13113cbcd17b4c3f29e35c54e30bc3bc31ae52 Mon Sep 17 00:00:00 2001
From: Jonavelle Cuerdo <cuerdojonavelle@gmail.com>
Date: Wed, 17 Dec 2025 14:39:52 -0500
Subject: [PATCH 1/4] Use metadata in 'bulk-update-embeddings'

Why these changes are being introduced:
* With TDA 3.8.0, we can now retrieve record metadata columns in
embeddings read methods. Filtering embeddings by `action="index"`
prevents any attempt to update documents that do not exist
in OpenSearch (`action="delete"`), which results in an API error..
This is important especially with the current state of tim.opensearch.bulk_update,
which will raise a BulkOperationError and cause the 'bulk_update_embeddings'
CLI command to exit early.

This also includes an additional change to also index embeddings
when performing a reindex.

How this addresses that need:
* Filter embeddings by action="index"
* Install latest version of timdex-dataset-api (latest commit)
* Update embeddings in fixtures/test/dataset to use 'embeddings_timestamp"

Side effects of this change:
* None

Relevant ticket(s):
* https://mitlibraries.atlassian.net/browse/USE-273
---
 Pipfile.lock                                  |   4 +--
 ...ef5-5629-4bc7-b743-7804a34f9593-0.parquet} | Bin 23387 -> 23415 bytes
 ...0473-7274-4ccd-bb90-f3c9bcc1801d-0.parquet | Bin 0 -> 7317 bytes
 tim/cli.py                                    |  28 ++++++++++++++----
 4 files changed, 25 insertions(+), 7 deletions(-)
 rename tests/fixtures/dataset/data/embeddings/year=2025/month=12/{day=10/629d15f4-84e4-4b32-92c1-1b1debd377fb-0.parquet => day=17/4a40cef5-5629-4bc7-b743-7804a34f9593-0.parquet} (94%)
 create mode 100644 tests/fixtures/dataset/data/embeddings/year=2025/month=12/day=17/8a790473-7274-4ccd-bb90-f3c9bcc1801d-0.parquet

diff --git a/Pipfile.lock b/Pipfile.lock
index c0e2065..b1b89c9 100644
--- a/Pipfile.lock
+++ b/Pipfile.lock
@@ -735,7 +735,7 @@
         },
         "timdex-dataset-api": {
             "git": "https://github.com/MITLibraries/timdex-dataset-api.git",
-            "ref": "184d87fd0647cae9e1db8488f2a3d88de96a7226"
+            "ref": "a1d8ad7662d8d864d1fb2b52376312c7c92e7398"
         },
         "typing-extensions": {
             "hashes": [
@@ -2107,4 +2107,4 @@
             "version": "==2.0.1"
         }
     }
-}
+}
\ No newline at end of file
diff --git a/tests/fixtures/dataset/data/embeddings/year=2025/month=12/day=10/629d15f4-84e4-4b32-92c1-1b1debd377fb-0.parquet b/tests/fixtures/dataset/data/embeddings/year=2025/month=12/day=17/4a40cef5-5629-4bc7-b743-7804a34f9593-0.parquet
similarity index 94%
rename from tests/fixtures/dataset/data/embeddings/year=2025/month=12/day=10/629d15f4-84e4-4b32-92c1-1b1debd377fb-0.parquet
rename to tests/fixtures/dataset/data/embeddings/year=2025/month=12/day=17/4a40cef5-5629-4bc7-b743-7804a34f9593-0.parquet
index 90b2f2a101f18a4d48b75e924c6ad358171afc2e..81e5d798970611592790c28ac2ab57a0793262d8 100644
GIT binary patch
delta 681
zcma)2U2D@|6iz}((}^JH8-@tpjfsOSow9a}jguwqI-NFIz7?cNyR1!9hpQI7aNeqy
zFL*8U<5r<C@Wwykr9VUjFMLnh;=r4oi|2ej=Q;0rf1dKcPWi82`0)D4kMKrK<lcNJ
zNriVGOVaw=Pm@aadEyC2WO8+h%#lk(B;vJsY1WySX45ja_+NEg5?bbk%bhIhwA<Z5
z@@SmyEeiL~6o%cbb2MyaheVL)<X3|HhhI1;E_^@#4w0k0bA`TiE{=aYf>3%dDrhRu
zg0C+<1{2^37yxxs0{{o}TP%P(31S5S8({S4(BCNPy_<?st<Z12UW}Woin40cLa@o(
zR1Y>qn>8x|eI2YF+iXd-fvZqq>i{+22qBll0Ki5FJ+&KpR@(Az5BwGDF!J0N?Rsq(
z12h0epe5+@=P-6-I={<e&2^1k=r`Gt5lhVmdpy8hgkwBpY6gbeJ8IOgJ`cTn&*IG5
zkIG>dm)&KgYymK}1hf$@n%W+EJ(wM!S$RD`Jaa%D^LXGs^~?Q~DOHfqe$#oF3ns9Q
z7cevbO&K*tVBGwoJj8U;*OptJnuh+tew?|zXk{mz2qPhv$oZ5!*pOn0<C<Wpe*wW#
Bz=r?;

delta 595
zcma)2Jxjx25KfB3L>ycqLBvfdD5R}Qt!T@oNuz<5rnWIjC$$=}sS1A7X6fqwE-pH|
zDG2@xe}VtNS*@e*y|Ey;c*EUuch5a{+~ApiedZq@cz^sod2we5T>V25M_QkfxOO&(
zE6uMs%A7nYPc6t}@`NnN!n7g|JBm1*p5&5$s!vj4$L6J{i>+jWm&ko%dhVZtZ;jc6
zbR(!}semZ1J|2J#a0&)MsY?#snM6q^*KU5)(C1Zk-XJfoKDuAcs%kbOidWzx;(7%k
zqRo|n+<L41h-&3*R0II6ct|}!8=ZzOHVZ?`49|VHlJd6GUCfIBWdIN3Cg#*Aq~`$W
z=p3h}gH@->D>O5T^hU~HqYPpSeGReJYi%zd>~)-@-9@YIW^x_>xDvJ)BS1B}U=P)Q
z%m@%<s#F{G>;~IvvMz%RX~^2J(z{{q)Ti(g`B(H)OwR+Y3*|a8=?TLrO{-xAPSE}O
m0PC|4n9h>BXoX%T*YiuUFr_kyn7_VG(#^bRiX2ykTKXH=qo_mx

diff --git a/tests/fixtures/dataset/data/embeddings/year=2025/month=12/day=17/8a790473-7274-4ccd-bb90-f3c9bcc1801d-0.parquet b/tests/fixtures/dataset/data/embeddings/year=2025/month=12/day=17/8a790473-7274-4ccd-bb90-f3c9bcc1801d-0.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..af9e013b4f6598fc2d8cd065fc59907bf8fcb125
GIT binary patch
literal 7317
zcmb_h3v?XSd7cp?Gi%L`XBX|Vve#zUvbGXv?R#hLorjTwN4EUJCbq${tPq^h?nqjD
zAM!qAo0NxoYFZLpLPA40Asp0$PiZL-V#tA%v^d2r=0KsG6H0U11SL2va3F0%=qqXZ
z&C2ovo0QXZZST(9$N&Dn|G(E;-R(3c#mv~s&9Ro45EX<)<ChA8P%Mm$H41sZwsQ6-
zOD%3QGbY5Mv3M_*YO#5-c}Y{~zVsuOCO&E+5&`{Css-OIqSzvU{nprZbL_kCTZRdp
z`qrp!=~-*UF|*7jxvZ1ZdDbwA#SP9!b1ZiOS(bgAeTk2?>H<>AF=b9lNr8~Uhb2fz
z!RYLFT<AXQ?E=)aSo<6`Ir;wT6FYm>%{!l>JU5d#{RGOVpU;e5fWvKK?8Z43Ylg0Z
zXq5@ST=TtZZah0tt?co0^%aXRxwY&!s$MZ$oA9bNKkJusm3*N*md#gk*-0ZiR`)-w
zC%k^u1xTGW^uLl!{JI1#pzo5{(m8b(*lkdEv{5YXuKKx3HGcsW+RmCzoc(eE(n73d
zGe`va5%_3fY)S8#90(#gOq%g5%zg(k{rq6;d;N3rJb1+y9*7CTXP4%^y1#T~iY&Jc
zn-In*(M{8~OuaX?bi~WIE?R3*+c0fP4Z?_{GuoSKwS-#h;?I~CA)HaevWRJM)5MNh
z-9k&YQf?T?3~o_k5^7_rUMPhYfr4Qh24f}%A6s;ULWtW~Wn-1YbYkj;C2}Rm@z6sy
z=gh*m&J3LqgNs&ZB6O$C9Mhqc5tCDk=yt3xvSb5u49m13rLJ?^0lUu5@Gim}X6d@k
z3@oIr&nuH&u@P$Yfju(}2Sl*TqJ}srcXl3VeI65r<=8fZA-JR4j+jgjhF)9^9Zj2n
zvPp>ry^Pjc$IC+?U||^y!yU`Ez;l1KQXXp^-fEZ@^oNcXv!JccG3WKHf*EyYJDkIm
z97YpQNX0;XEV3L}#&QTT93tMC%y~t>ocF4c^ajIrumGl~mdQY8VtgXJD8&t%a^0pz
zdO%#1fDgS9uhw%hv@#si00YB<&AN^%m?}gbvvuYW9fH^<3~h4pB_SIcVmft*rE?uh
z>9*C?DcANDLrI%5({eb0mx#?xrzf3uBbN~dZk7c%!9`rWJ2B~13*JaEoQAbbKp@P|
ziS1DLYr^i<m74B2j_LNabFMqoO0R04I9MOAl*3CHN6;W77_5f~5{DGOTCMaf>(*fd
z9bk%(5tj(2xGGWh#~}V>XwudZ1B6*{FI)!`rIbq5%Z2Yw!AS;XjM?rRf;sSD0@z>m
zss}^A2}47A3N8pW%35!bMhme%0yh#kKqm|#Eq<jPQCX>vNBWZBidf;cS8aVw+h3{H
z>WCj5al=3ZTy>iMFS%B)(g&Ds!(uQbXYer__@Vm?p%%Up4$?W6o2KJ%UEHs9b}|#z
z<S;mVhdFqK+2Uelk2mRs7sKJif(xJ%QiP}5<b4f4@(yFrm6)&s^2e~!|KJaX8#aRk
zmWh?{IlLzJ3AxaqV0gc1r!X_)>A&(#VI;JuT{j#XyPi@k-7>{%h0aKa3_9-3noc=4
zMW&5bswFR*WKaR2iIk%VYT|EcFO=*4n4kTr80)sVZV=mo;kY;?)x1IiDw~kPCde9O
zrJf!V>N}=v>X;_72#IPk-Jt$+sqsRsUa1yx-o`iks0~wq5a5Iu7I&t;+@7nHC;e*Z
zz7)kCotlmgU)$N6g#6)tad9`Xkdm;j#mJEbQm%6N$v%pVv`o-3OoL&UIP4wmir5%7
zDF*?l!A*XsO{?Xqg^BRwGNhbka|leluLx}Vr?;<dr^q&oP5+(d7!I=!{X{7@N+a1T
zBx;EzjY0!MtVy;pg?Ok}3nPt}wxp<O0iYaerk|A!;<&oxO@u!!!!ce*!a%$OuEa%3
z&96>IuD4AcISiPxbi+0sL;U?C%jIIByv5WghxAjw-43*1wmz^UQT4&XpDcW>j{-a?
zF?0)>Aw+d?tr9dfJ+W?B#xRjrP&nP!jjS)A237&C*-I5ph?D+_K*d#pGw`0&*}3YU
z#ndt#G?0S`$$=-e&d!NKHGC68Ks%fvWdk78TaQV(0!?U)f&6rvMofn=@g}#2zO4kn
z;f`rKacsS7GSS(Y^C#+s!%u1ui#Xyh)B9!am|gEBE4fC=FV{UiBea`<4@SI{M0t9o
zJ@1ctjbeBP5N5$(;A()w#<wN^fOzc0(LQR}w!z>k+p-*_Z2AW$-U$YEm<}llxWMxC
zu0*9#FIFmhd#)!yT4I41kc_kAUj{}OB7HK|t*P}g@*(}E)SD`0Zhu{-26Z1!AwNJF
zt`|R*tooyV6^%o7mkn891cV=Q2ly>6mGUE@19$=2=!QX2aIE8NB){xNZXc7VKK;8%
zhTt+BY9H#5Q85;l<ZcQHP`HrCJj!J8Jz-^d5Z2O7LwrmDL|ccxr;T{^+<0h70yuz7
zG1DgmUA;z*{)a3!gh{%=rl<!5;+YJwMRajkI#4<sMQw2y;z_3{t|ob@Qz|#Iy#NQ~
zIR#IMg3|SkW9jhA60v5!m4c~kY9mFa-b$b!F4qh7?B<{tGC(_w1<CpGF84d}D`gwS
z7qlG{lA=nHlYa>t`jHz63o3z$g@AGry?DNytNKg2=BE&i3}|g}*OnoU_%mU0;ZQ(E
zN4T3d#B^-aawp+a{m)tupcYVt0Z%PJaO-OdQoFY8`$<^bwh`FTTV%<$O0UZZ-pA*=
zLN@_O0{KzWN%YS{%D7h>@A`s<%EyShMowTK@|mXNnm6I)#y5VV9qj_)j=FQcW`d;U
z$EJn^Z`<x}V2TYd!FFca6zITi85wnOC<(|%nM63~GSQj%bFEsb*Zuf5nGi^?+Z@%x
z*7?+rG$d7{cK9_=Lg*s^IXBe%=BJbGNS;ceT==3yY;Q#Zz-EEfRHuv$J&*Lk?xv{&
zxlzLaOjLa<q>L8)V&RX2Da5=9|2k-ls40F)u8i!t>xG$nl294(W7x;868E<Q02<Zu
z#*J|2bhiSuGR3umbNrt8=C&n<+zuZ)Iz3rZ;2_qzR6|}B%Gn*=7F>lU3jjmai$mg5
zQlWhC`d+L?c|om1@8A~2DXmub>W#>)%K(4Kd1@maQfWiWdF65?8=qPsIPM=vI#ZvP
z#ZPFtO0iMO?o&*MIclS=7r8=A(+YZ-JsaB;G(CLg380;hnx>0GL6@}ghXPR_S7e-p
z)Rz^>X18a7LD{W>nSQ#R={kXUskU?(au@x2=~XcW=+kw^XX+9mVhuF%Z<Q6W(bcbp
z5{P=3j79I2H7}R*YvFg%HX%`U1L;Z87M|o&M+!$JVn%OJ1QMMJNrXlJG?c>Gz$^K6
zzZ!NCA?PYm3s56%^2jd~KUef#I@pJC#6Hpj32GXqBYs<$h;LNnAfkRtNda9@Kv0ss
zq&!~P-;<S*6zRWEID;|vDCtqDSc%x^7&veXIAsCgkAEG?>Q7vRFhi#dF?FN`QbTN<
zydv_1dW#I>bc8kZ9mL3fE=vhV`w3)*^SMo?xX`QK85Z`1@98E;K8ipxxg+);*?mIW
zaleM1i<s&!6<|E)16K)GL{_Qy%PFKZ2T0n^p@-xk7Jdu?QQv^4LGMPL=rUESqaE5C
z8<Lng<0a8RppPOXx>{DeYSqg=49tgH(JR0r2;KO1WF(Bg@5Pz-(=&($?}LGxq{hV3
zSCsVMcLTjp&=Imo;JZT_@OYxJ<$WMMnpVI%z|ORzdlzUE0NKKpXA*EPsux!?GJseg
z$OSdN?iKdE+KX}s)Uz3yUAFLFLo!H-H-~>N7&G72(5yR5U84#t`uTA5-<9m+hZKQN
zj|p^ozI(Ye;Vs;)BmqYp0nVm>BvXeUIvl83C~`+K40Y#%6rj;U_h*o3BJ8D?E>lr4
z4T?q!=7FMWQBx4fztQuAVgOoa#sxa{bs&7`#!^dUX$sH}6hXAPHL<U~HsR+AquF~~
zZ&FbAS?eFR_5+{F`Gvc}@acVR>SiT@&eJp~J@k8#VTJ0ZrQQ?PaK5m(e=}<CxL<th
zktART<;XGyvNB3w;omObgi4DVW2j17!UXc7yB5>J$anAUMa@Q!ic%8<TvZjHLd#ek
z4SiSLG#{nIizpTVF@7Mj<PA0ZB?X~ox3)++@lLb`EAy4`YBVtF?P?Nwpb}Y5^fE=N
zd!=XFfyxLO^-fhoZDmI5l~SX|A4`BY!p=}Ds*C~<?Uv=be_+e!k|^;=aB5CJy$o$5
z2~KM@KfG9?=CSuQoDCd4^Ea)N%Lm%pgYF;A!_DfBkVNh1^Qu-X<oxo&t2Ke7|1o8w
zQwGp+uD+o8Gwl<VTA^NfJf=qH3$}ZQj8lgAfZ88Q)N<o~zERxv3i=E55b7<eq&w3`
zTRWjwxw3Fyo4R{}z>XYk$Dsp}{_bZ5KX#dnz*Jw9lW0)@ji|i0sj&rH646{}8JvM;
z2VrZS6i`2V4oc{e-cF$^ny69Kh(0Ex_u2J!ieN7ePK-{;60^Hj%loT^$c@wQq)^~F
z>M*n1pQ!_3xsfkqJ)CPS8~|_-#L>w}yFoC0Dtu5w%5pvZO*D+h)akn9<x6h>-_hP7
z#cia5ZT`_+Qlos=YKd@fu}lf|_DX0Reh|{UTCFhl?zM=h@2ktuL2>{z&YIH=%|B50
z%b5oOZ(Lm!hFvVgS3WE-Hx8pz@(Uxf`c+NFN$1#`f+fDLMr9#iS|@JnMnR=$BvCf_
zkpqCIOc5wFb5IK!d`rDGEEuKRaL|mu*(P2lP<K$qv0ODS>3Ucmtu$hv4h{l<Pqaqp
z1{@<D6DG4nLlbAx^i~=8dF;`4D30ryg<eR(Dcn&1_7Zf1#<4#+{-%&$cwIQW8sR#2
zi9pkTEjgz8w(K1f3ekOS0IS^Psrjm`NQT+^J7~VqX-8Lvr#1+sm;PlLEQK0{GXPZ&
zMkPW|YI45Xh<tuE&Pi|^+>ApI{#ZAie^-JFj&m;TI-cP|VqOqdB`>B4F{P^|q2X+o
z_Ya)9Z`obQcL|%5!F5S6Ij`C%pEcBip-3<^dx5dLGCErG>$o++_$7X6#LoveP`g3V
zM}aL(bhU53Ba-X|wNvyKoLN+=p!A*1D|%)Q2T0tvj2%2R%^~HP$w^#)RjRngnYXzQ
zbh(s(+iLX~ZX{*MaWM*N(0iZK?#w=P8Q3YrTRM^*cVvWQOzGglJgwt=3&dLDE%6V;
z_<xI;8#40$^|G|17q>~cfC?_2H2n0%mv_wT$TY8-&Srng#!aTrwZva+$$Zd~xveE&
z`>}33KlZ)o`2uY?N1&Ed_Bh|eP1)`i<M)Y~|CBPXO37b&V|pH+p0m}KPI0+Vyfpbd
zBgUT-GxyKSyw{pMb5DBa=_Bt^e>LAZZ+|!hn>_#NqWBvZWnPdo-?%ui`>%V+`TucC
zdY(4%o6qsqoIV#wo2E*Sit%rXnOE8}Uu+9hI{TvZ!?&jAvx=X?N}hF@eO?ka6K^iz
znp{5<AI<z&$vm$l%@*N2x;+^Tt_$v3@i{vO&Z0b<rcGiG?u>tHP3EDknMcD=;e2xO
zX9rKq8GkSzZ^_8<H?|5L$%Nu=-~P#+D`$}|@xZ(l!H2qQT-V)(f5GQ!*WK^-kGbny
zcg$Tg)*S2acik<)6Bcd^e(^r{i??9W+Dbs7>JGT>`fZqYcLeVY5AM}XB6jc?Ot=Gk
zu2``mcpVtPY7lL*ZaxODvy1wRBRe-0hu3eP7%AmPNAykmhId{=n)7Re{f@n2a>a^Z
z_1gaC{>tgCfef=k%|5LktnVrf?j2rVYz*TzIG{6PG{tvwTdfr<tn23FS=-#%zdzW2
zMxLf=`X}<BgYoqNY1jw9%}>Ak$<zDSH8-87Pr&P3eR9V3gS&>d7l)0#u6s>$|E|Cl
z!NUISO}b5^HSN`Wb+-mT!#inVc<1)L-cDW~(&^;z`oWs(UWc8|#<~4##+w^3e#(^1
zSNFQ$=bZfpY`b&t#{Bwq^<2}RO}|6fCTpzTIO~e0>-z@>2e!Dyt>6-vXWPc6(KiP~
zPlpbS7L0D#L`RCwLD+dRSK1EW4VQBIVAFSqAH)&tw|_Q%n%l1pzSn+O|6Rf8S^78a
zohuJEhV&wgw#E+hg}ry2^2dfCZt>Z6X4Akae>C+u<Ig71e&X$<dxwX{M{}jYJtM}Z
zy|Z*rlW49N5-FS%akE;_<|ZbxI1N?yr-EyWLZzJ2_2pza>9_-3!3F;cyJSJwivN)O
EAIV$YrT_o{

literal 0
HcmV?d00001

diff --git a/tim/cli.py b/tim/cli.py
index a97f4c6..998805e 100644
--- a/tim/cli.py
+++ b/tim/cli.py
@@ -385,6 +385,7 @@ def bulk_update_embeddings(
             "embedding_object",
         ],
         run_id=run_id,
+        action="index",
     )
     embeddings_to_index = helpers.format_embeddings(embeddings)
 
@@ -454,12 +455,11 @@ def reindex_source(
         tim_os.get_index_aliases(client, index),
     )
 
-    # perform bulk indexing of current records from source
-    index_results = {"created": 0, "updated": 0, "errors": 0, "total": 0}
-
+    # reindex current records from source
     td = TIMDEXDataset(location=dataset_path)
 
     # bulk index records
+    index_results = {"created": 0, "updated": 0, "errors": 0, "total": 0}
     records_to_index = td.read_transformed_records_iter(
         table="current_records",
         source=source,
@@ -468,7 +468,25 @@ def reindex_source(
     try:
         index_results.update(tim_os.bulk_index(client, index, records_to_index))
     except BulkIndexingError as exception:
-        logger.info(f"Bulk indexing failed: {exception}")
+        logger.error(f"Bulk indexing failed: {exception}")  # noqa: TRY400
+
+    # bulk index embeddings
+    update_results = {"updated": 0, "errors": 0, "total": 0}
+    embeddings = td.embeddings.read_dicts_iter(
+        table="current_embeddings",
+        columns=[
+            "timdex_record_id",
+            "embedding_strategy",
+            "embedding_object",
+        ],
+        source=source,
+        action="index",
+    )
+    embeddings_to_index = helpers.format_embeddings(embeddings)
+    try:
+        update_results.update(tim_os.bulk_update(client, index, embeddings_to_index))
+    except BulkOperationError as exception:
+        logger.error(f"Bulk update with embeddings failed: {exception}")  # noqa: TRY400
 
-    summary_results = {"index": index_results}
+    summary_results = {"index": index_results, "update": update_results}
     logger.info(f"Reindex source complete: {json.dumps(summary_results)}")

From 10a887fac4ebb5af233f2390fd1f109855de861e Mon Sep 17 00:00:00 2001
From: Jonavelle Cuerdo <cuerdojonavelle@gmail.com>
Date: Wed, 17 Dec 2025 16:20:01 -0500
Subject: [PATCH 2/4] Fix unit test

---
 tests/test_cli.py | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/tests/test_cli.py b/tests/test_cli.py
index 61dbc4f..07a6546 100644
--- a/tests/test_cli.py
+++ b/tests/test_cli.py
@@ -344,9 +344,11 @@ def test_bulk_update_embeddings_exit_bulk_operation_error(
 @patch("tim.opensearch.create_index")
 @patch("tim.opensearch.promote_index")
 @patch("tim.opensearch.get_index_aliases")
+@patch("tim.opensearch.bulk_update")
 @patch("tim.opensearch.bulk_index")
 def test_reindex_source_success(
     mock_bulk_index,
+    mock_bulk_update,
     mock_get_index_aliases,
     mock_promote_index,
     mock_create_index,
@@ -362,6 +364,11 @@ def test_reindex_source_success(
         "errors": 0,
         "total": 1000,
     }
+    mock_bulk_update.return_value = {
+        "updated": 10,
+        "errors": 0,
+        "total": 10,
+    }
 
     result = runner.invoke(
         main,

From fa936ed6e4032507fe424b0cf433e8f7daffaf8d Mon Sep 17 00:00:00 2001
From: Jonavelle Cuerdo <cuerdojonavelle@gmail.com>
Date: Wed, 17 Dec 2025 16:22:05 -0500
Subject: [PATCH 3/4] Update dependencies

---
 Pipfile.lock | 44 ++++++++++++++++++++++----------------------
 1 file changed, 22 insertions(+), 22 deletions(-)

diff --git a/Pipfile.lock b/Pipfile.lock
index b1b89c9..20c0553 100644
--- a/Pipfile.lock
+++ b/Pipfile.lock
@@ -26,20 +26,20 @@
         },
         "boto3": {
             "hashes": [
-                "sha256:cdd4cc3e5bb08ed8a0c5cc77eca78f98f0239521de0991f14e44b788b0c639b2",
-                "sha256:d21d22af9aeb1bad8e9b670a221d6534c0120f7e7baf523dafaca83f1f5c3f90"
+                "sha256:649b134d25b278c24fcc8b3f94519de3884283b7848dc32f42b0ffdd9d19ce99",
+                "sha256:8112e1beb5978bb455ea4b41a9ef26fc408f6340d8ff69ef93dded4f80fd53e9"
             ],
             "index": "pypi",
             "markers": "python_version >= '3.9'",
-            "version": "==1.42.9"
+            "version": "==1.42.12"
         },
         "botocore": {
             "hashes": [
-                "sha256:74f69bfd116cc7c8215481284957eecdb48580e071dd50cb8c64356a866abd8c",
-                "sha256:f99ba2ca34e24c4ebec150376c815646970753c032eb84f230874b2975a185a8"
+                "sha256:1f9f63c3d6bb1f768519da30d6018706443c5d8af5472274d183a4945f3d81f8",
+                "sha256:4f163880350f6d831857ce5d023875b7c6534be862e5affd9fcf82b8d1ab3537"
             ],
             "markers": "python_version >= '3.9'",
-            "version": "==1.42.9"
+            "version": "==1.42.12"
         },
         "certifi": {
             "hashes": [
@@ -644,12 +644,12 @@
         },
         "sentry-sdk": {
             "hashes": [
-                "sha256:8218891d5e41b4ea8d61d2aed62ed10c80e39d9f2959d6f939efbf056857e050",
-                "sha256:d72f8c61025b7d1d9e52510d03a6247b280094a327dd900d987717a4fce93412"
+                "sha256:5213190977ff7fdff8a58b722fb807f8d5524a80488626ebeda1b5676c0c1473",
+                "sha256:6b12ac256769d41825d9b7518444e57fa35b5642df4c7c5e322af4d2c8721172"
             ],
             "index": "pypi",
             "markers": "python_version >= '3.6'",
-            "version": "==2.47.0"
+            "version": "==2.48.0"
         },
         "six": {
             "hashes": [
@@ -927,20 +927,20 @@
         },
         "boto3-stubs": {
             "hashes": [
-                "sha256:b132be8260eb56010b47499658cf6c485f99ea0190969e7d1adb74c505c83e68",
-                "sha256:d9b108fb3e0af33fedb0a9f0e214bc1d474a6e092a4b8755e88ccc847dc9c624"
+                "sha256:18a9a970fd9d8867558f091608f2d9944d6f55143d23cb11785f1c143d91aa54",
+                "sha256:f69f12c884519a62a6e6d7f296932d4cdc14f47ea987b1eb7b226337f1127d13"
             ],
             "index": "pypi",
             "markers": "python_version >= '3.9'",
-            "version": "==1.42.9"
+            "version": "==1.42.12"
         },
         "botocore-stubs": {
             "hashes": [
-                "sha256:92fdd2a1d911355166da3e30e9bb9b1803f7e2caec0d913f5fad3a920352ce6d",
-                "sha256:9f8b652549d4f727aa69e09d462d18e54a1bd10f3dbb593da56d5d0aafe9756e"
+                "sha256:788ecde81894f149bf210286f82ff3e49b97ce736a2ecb89f210be3f0374d35e",
+                "sha256:c2bac920d8c302e15e6bec9593daeaf04777714f8309fe052fd24cf28eb85a76"
             ],
             "markers": "python_version >= '3.9'",
-            "version": "==1.42.9"
+            "version": "==1.42.11"
         },
         "cachecontrol": {
             "extras": [
@@ -1253,11 +1253,11 @@
         },
         "filelock": {
             "hashes": [
-                "sha256:339b4732ffda5cd79b13f4e2711a31b0365ce445d95d243bb996273d072546a2",
-                "sha256:711e943b4ec6be42e1d4e6690b48dc175c822967466bb31c0c293f34334c13f4"
+                "sha256:15d9e9a67306188a44baa72f569d2bfd803076269365fdea0934385da4dc361a",
+                "sha256:b8360948b351b80f420878d8516519a2204b07aefcdcfd24912a5d33127f188c"
             ],
             "markers": "python_version >= '3.10'",
-            "version": "==3.20.0"
+            "version": "==3.20.1"
         },
         "freezegun": {
             "hashes": [
@@ -1651,12 +1651,12 @@
         },
         "pre-commit": {
             "hashes": [
-                "sha256:25e2ce09595174d9c97860a95609f9f852c0614ba602de3561e267547f2335e1",
-                "sha256:dc5a065e932b19fc1d4c653c6939068fe54325af8e741e74e88db4d28a4dd66b"
+                "sha256:3b3afd891e97337708c1674210f8eba659b52a38ea5f822ff142d10786221f77",
+                "sha256:eb545fcff725875197837263e977ea257a402056661f09dae08e4b149b030a61"
             ],
             "index": "pypi",
             "markers": "python_version >= '3.10'",
-            "version": "==4.5.0"
+            "version": "==4.5.1"
         },
         "prompt-toolkit": {
             "hashes": [
@@ -2107,4 +2107,4 @@
             "version": "==2.0.1"
         }
     }
-}
\ No newline at end of file
+}

From e0107bfc568f040eecc919b305128137fea1bd30 Mon Sep 17 00:00:00 2001
From: Jonavelle Cuerdo <cuerdojonavelle@gmail.com>
Date: Thu, 18 Dec 2025 09:42:18 -0500
Subject: [PATCH 4/4] Add docstring

---
 tim/cli.py | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/tim/cli.py b/tim/cli.py
index 998805e..58219bd 100644
--- a/tim/cli.py
+++ b/tim/cli.py
@@ -430,11 +430,10 @@ def reindex_source(
 
     This CLI command performs the following:
         1. creates a new index for the source
-        2. promotes this index as the primary for the source alias, and added to any other
+        2. promotes this index as the primary for the source alias and add to any other
         aliases passed (e.g. 'timdex')
-        3. uses the TDA library to yield only current records from the parquet dataset
-        for the source
-        4. bulk index these records to the new Opensearch index
+        3. bulk index current records from the parquet dataset to the index
+        4. bulk update current embeddings (if any) from the parquet dataset to the index
 
     The net effect is a full refresh for a source in Opensearch, ensuring only current,
     non-deleted versions of records are used from the parquet dataset.