From 65b5df3a431dd4b5c9189c7d9ef59663e7871fe1 Mon Sep 17 00:00:00 2001
From: bolade <babawale030@gmail.com>
Date: Tue, 2 Sep 2025 12:22:50 +0100
Subject: [PATCH] Add CompanyTable model and refactor query handling; update
 requirements for new dependencies

---
 .gitignore                                    |   1 +
 app/__pycache__/main.cpython-312.pyc          | Bin 2312 -> 2264 bytes
 .../pydantic_schemas.cpython-312.pyc          | Bin 1640 -> 1640 bytes
 app/api/__init__.py                           |   0
 app/api/__pycache__/__init__.cpython-312.pyc  | Bin 0 -> 168 bytes
 app/api/__pycache__/investors.cpython-312.pyc | Bin 0 -> 481 bytes
 app/api/companies.py                          |   8 +
 app/api/investors.py                          |   8 +
 app/db/__pycache__/tables.cpython-312.pyc     | Bin 1299 -> 2139 bytes
 app/db/models.py                              | 103 +++++++++++
 app/db/tables.py                              |  23 ---
 app/main.py                                   |  24 ++-
 .../langgraph_agent.cpython-312.pyc           | Bin 7279 -> 7268 bytes
 .../__pycache__/openrouter.cpython-312.pyc    | Bin 8681 -> 8681 bytes
 .../__pycache__/querying.cpython-312.pyc      | Bin 0 -> 3970 bytes
 app/services/langgraph_agent.py               | 162 ------------------
 app/services/openrouter.py                    |   3 +-
 app/services/querying.py                      |  46 +++--
 requirements.txt                              |   4 -
 19 files changed, 166 insertions(+), 216 deletions(-)
 create mode 100644 app/api/__init__.py
 create mode 100644 app/api/__pycache__/__init__.cpython-312.pyc
 create mode 100644 app/api/__pycache__/investors.cpython-312.pyc
 create mode 100644 app/api/companies.py
 create mode 100644 app/api/investors.py
 create mode 100644 app/db/models.py
 delete mode 100644 app/db/tables.py
 create mode 100644 app/services/__pycache__/querying.cpython-312.pyc

diff --git a/.gitignore b/.gitignore
index 698ec62..2ae4ed8 100644
--- a/.gitignore
+++ b/.gitignore
@@ -12,3 +12,4 @@
 
 /*.db
 
+/*.cypython-*
\ No newline at end of file
diff --git a/app/__pycache__/main.cpython-312.pyc b/app/__pycache__/main.cpython-312.pyc
index 9cbd458af81d89c7020e6ba300163b1926a0d2e5..6d7f7be4f30b04b061da4f639c9e031a8a6d3e58 100644
GIT binary patch
delta 1004
zcmYjPO=uHA6rS1HY&Ji;`PsHHwzdCgu&uubi`s(+MbM&%6!zefW~Oawb`xitmVlCA
zdk_!OvIjkS5}^u$;K5VxdPr%ZaZAvG2fYP*v7VgSnA!*Xe!hM0&G+8SlXxwrd{UGk
zVDjwahuPAqk}N-k06Yg6V3-3sw?Yn@f-b-&hIEl(DCaXJUGgw7=Qm|tHUoOV4C+A-
z^W{QjSPy%?lv7MqS0NApk=OWHO!Nqe%=I<>!<ryruHPA#j*1YF$m**R7+KxwD+h3J
zjq&0#4RMG_I81`mB371t2OZm4$5b=tu*Pbu(yNP*4ptXk77ywY<Oq)4KLG&k^rlYg
zNyaoecm98ut1TDDv4-1l`<sq6v57%Yf+d(r{N&+83c7);wLt9oJhfS%;3)%>0<kc$
zG9_mno{K?Oy<H@<bc^OQ#J2O)85S-=)tMFUcEwl^m3_9tR0VtsYb%r0P<K`Arjhqw
zgk31b*$=8Y>ta_b!CV(o+s|xoQ)rZ9fkK+v#K_nSbdbfG<MVy)bL9s`Mk|}QG6Dlz
zf|#?!zP%d`lE%udRXsvjc!;v_4tR?eVG2#qc9ul>ktu-&m~oXta{z<&rKOZeJ!(Td
zt-qCP^CfX=O=1x*)i_uXWlIPA5!ff0`T>5Qu7K~+W3SI{MMo>wYieR!9okTbUe0c+
zrz+QK;r5lqYN)3w_UwC&ANLPB`_h%cPTI-JU6d|lDWAuc3TX?cZJOkP4^NHaDd(R5
zLOEs=x{zgCAN7c2tr>3=%1?LItd+?XF)?VqI8Uf6{j+F>%u{|bu2{(1^WI8b#2k;k
zD;tKDHi==l{vzw0r<RLo(W3ksX@o0&Cx7HiX8tBF=Eyj`%wqnQ?QNzS2tv360(Fqw
z0a_jOd<7#LV5APt)WPtsq%37Cs0PqBh;M-S%GkCxu%QiXfx(9Cq~*!pO{g^l1bbhY
X4ZzGh`xiHR2ncN+?$qVuB4_azf79&(

delta 1096
zcmY*Y-D@0G6u)<7_A9%y+1X9HTieu@py>t^X{^CwX|-VcU@TaXl0g_IdvB7B^X0j-
z2@6UVn+JWeWgfKufshsy@n7gmE2LB~+)`?=(E1jvQ0tTD&LpCTx#!I9oO|v)=lo{2
ztuN~4ZPOegDAyl<zx?Ah^Wb_w3Hh2Z!l*~=%u-uJta-X`*oL38b5!|tFYg!Zf?u?Y
ze#tJSKEoUG%XT?^&v_N!v`xQiS5ut#hW(mdqeLUTu&M4uu`NWoeMAYT*0ozxsQx^*
zUSva?Sd5ZI%F5he6<%62n7Q$9f<4Num3Fr0znWEf^~Ug~mif}$8aFZ5+A#Jnvk&kl
ztABi)5IB^!+p=3&OlxH(%bNS&fekjoM%jTI%}pcoQ_`n>+HU=>;KjBkO}5}L9`S(j
zpwpLTH|WL=b7ObG6+DsjwNsj`yx-%{zr;W3aS=;v4~ZxY1fT1QxJ_ka&JC6>fEz8R
z(YKd)5GO~pH)$=gwGR(B@U|K*VWJw4Un`BPm$%AeTgDhzKit-K=({5rfKgPk&<TSW
z)1}51WhsKN!-WU|4yqspQ9ye%6R0a+DRMz@rz2M3WyGV&Gvm$-)>H+a)#&&k<DM))
z)<I`)z_^zX9oH3HPd`c?<xZ`y?9EAq_GO>W&-73A56De@jkdLm(8O?ChcX(eLoblU
zbUfWKXd94rgB%f=bJZwh9&71&u~5gW+Ox`sF$MZgH(15Nh9Ka~G{(nq3H5nAA%6|X
zqsoilp8L7->foJiv;Nimm-AQGerZnJZBE=Xj}Ojom#x8j&(mW+pL`{`lRr290*qm<
z%tvm(Tmgp^D7jHI%@*(h)o?OXcw>D;`Jz5^gSgvqL}!`%xV?m7Q2&-Ztm1;hY8MHZ
z_K@Et#jle#l=clL)INQQ$Djz7Mih!T<p5M4vgkO0>vPAEg&v*~LI9KK2lXYuoTaoP
z@bBVjB&g@{GjJB6`gz1psCIQq>3vfAjWiyTnTKTRE}42jPCO=WJs_ud^5&K9K;I$e
v#_(=FnJr#^X>0uCpmmSFzN_i9{*Av&(0u3pq0CdAXtSw5xmlbr)Ybn0YF!c6

diff --git a/app/__pycache__/pydantic_schemas.cpython-312.pyc b/app/__pycache__/pydantic_schemas.cpython-312.pyc
index 49fa62c1f046168ca3768de4a7821f5c96636635..761fc876e847fe16ff24ae90e532fabf0d47fd75 100644
GIT binary patch
delta 22
ccmaFC^MZ%>G%qg~0}$+%+?3I?kvEzR08N$#egFUf

delta 22
ccmaFC^MZ%>G%qg~0}%9{*_hG2kvEzR08xYo0RR91

diff --git a/app/api/__init__.py b/app/api/__init__.py
new file mode 100644
index 0000000..e69de29
diff --git a/app/api/__pycache__/__init__.cpython-312.pyc b/app/api/__pycache__/__init__.cpython-312.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..707593a0b1e3feed607800fb0dee58913358f498
GIT binary patch
literal 168
zcmX@j%ge<81j~1A%K*`jK?FMZ%mNgd&QQsq$>_I|p@<2{`wUX^%UwSsKQ~oBKc}=j
zu{bd=H&fpwKe;qFHLs*tKRmxETi@5)ML#jGBtI{{JhLb@ttb(wI<cSth%@!$<1_Oz
nOXB183Mzkb*yQG?l;)(`6|n-%Wd!145aS~=BO_xGGmr%UNS!N0

literal 0
HcmV?d00001

diff --git a/app/api/__pycache__/investors.cpython-312.pyc b/app/api/__pycache__/investors.cpython-312.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..837640b8f56a4ddf65529594327a4199c7d69a33
GIT binary patch
literal 481
zcmYLFy-EW?5T3cqzb2v}Vr<e_B$3<;h#;g;j6t*zQW%!pCOOUBhTS_uQbiCHd;q~W
zkn|S5f=$5gK!{k`Dd|$Vd*)|g{+O9>zL~dVG67O9USFC|(toLBCi*lOmt^n=CYaa)
z8!L#Q0mZ?zT+yg@8J!I+yr~WFKQr}Pn0JS|p=|hD6|(5Z_fHN_ndeg8OBj}OO@+(2
zs4YN1t^!a@#)_z9fu?5R8D#acAGN6vq(OVDB}=#&R?o~Ko`FZykW`o)7)@qVgSEU4
z5yY`AqsVHWO&wPl=Zs$&$45njI4*Ol9gEXCC$gKg+fuCst3tc$r#K~Mb-sniC45+_
zNJ-O&(enCmBR5>i4?}sbUQ*76{gpZqu0$$ud8p+yxVpxVsEl%7Z%{XY3HSy>JSH0^
zv0a;)UW@MWq)gQ(MNXO@9T7iZ`!l}wu+@)@FnX81N#B>=ar0M0+!^bjrGqHI`U9K*
Bb9Mj#

literal 0
HcmV?d00001

diff --git a/app/api/companies.py b/app/api/companies.py
new file mode 100644
index 0000000..d718c44
--- /dev/null
+++ b/app/api/companies.py
@@ -0,0 +1,8 @@
+from fastapi.routing import apirouter
+
+router = apirouter()
+
+@router.get("/companies")
+def read_companies():
+    return {"message": "list of companies"}
+
diff --git a/app/api/investors.py b/app/api/investors.py
new file mode 100644
index 0000000..42d5b39
--- /dev/null
+++ b/app/api/investors.py
@@ -0,0 +1,8 @@
+from fastapi.routing import apirouter
+
+router = apirouter()
+
+@router.get("/investors")
+def read_investors():
+    return {"message": "list of investors"}
+
diff --git a/app/db/__pycache__/tables.cpython-312.pyc b/app/db/__pycache__/tables.cpython-312.pyc
index 8e6b31ce548b07a62420e563ee9e31229f4d306c..18d72cf2655e4fc4da75488271699723da0d8185 100644
GIT binary patch
delta 747
zcmbQtbz6Y%G%qg~0}%Y!xh+GCbt0bxW6VT#cPS<YcZL+U7KRk|RQA=(AO#EzQJj^G
znj9O8SeT4fGTq`XD9X%DEUJvpPOZGfnwgi9TH&^m@fJs3X--aJQcfz6&7P8)mROoo
zvXZGtaPtbLyNn8|Kx2wHfkZOMKp0?$GCsFW+~dH=KJl-5IBN>i8rIcJ5Q!*mCWcg&
zRCW|G9<Ug1Dqofi%p3%rC68dkSU{BulMR`qC$D1UG8H6Kx0D{q7OK_@RdQ%@-r{l2
z&n-yIs|*2!&MnSlFfTK;IHZUh=tQ?7J|I~n03w8dgrBBxkqC$*3L?ZnggA%*#YGV-
zh>!pgY#>4sNZjH8MnP$DNm1o3j-34D#FEVXyjwhJ`K5U&zNsnkm8pqEx7f>5lZrD-
zQj1tXx<ElxqyQomL4*>BPzDhyAOajyY9N*di~w0vyngaAW>tND(Fth_O0V$iHhA1b
zW(!NsP+#G4McA;x=O&-Xgs=rESNOCW+$XEC#Pfnp1R18uRU|)o5{pU}Yf6${N>ULh
z`oW0;Vil_Ozc_4i^HWN5QtgTifa(}QZqNi0AD9^#8E-T2KWDJG%wX}9LF@|~KO@@&
VmoE%J>WdN|kZtmn0Z4&;0svFPmx}-Z

delta 204
zcmcaDFqw<*G%qg~0}!-7-<WZeX(FElquWGvca};<O}334EKEk4Ot-iTiZXK(iz?%@
zQ!8(=X6B`&R=8<0-r~qB&B;kj%1H&X*;7)}5=(PRG?|L{Ht%D)%gCz;G_!~kNN92v
zNl!k@t`f_cSWuvsk_1FW>@dk91rQ6YRtd=Z#bJ}1pHiBWYFDHR<T3(ru`H1Iz|6?V
ec$<O$IfKPz28*W*QeW8k8QCVdd}RPqVC?`$3pFPI

diff --git a/app/db/models.py b/app/db/models.py
new file mode 100644
index 0000000..27c09ac
--- /dev/null
+++ b/app/db/models.py
@@ -0,0 +1,103 @@
+import datetime
+import enum
+
+from sqlalchemy import Column, DateTime, ForeignKey, Integer, String, Table, Text
+from sqlalchemy.orm import relationship
+from sqlalchemy.types import Enum
+
+from db.db import Base
+
+
+class InvestmentStage(enum.Enum):
+    SEED = "seed"
+    SERIES_A = "series_a"
+    SERIES_B = "series_b"
+    SERIES_C = "series_c"
+    GROWTH = "growth"
+    LATE_STAGE = "late_stage"
+
+
+# Association table for many-to-many relationship between investors and companies
+investor_company_association = Table(
+    "investor_companies",
+    Base.metadata,
+    Column("investor_id", Integer, ForeignKey("investors.id")),
+    Column("company_id", Integer, ForeignKey("companies.id")),
+)
+
+
+# Association table for investor-sector many-to-many
+investor_sector_association = Table(
+    "investor_sectors",
+    Base.metadata,
+    Column("investor_id", Integer, ForeignKey("investors.id")),
+    Column("sector_id", Integer, ForeignKey("sectors.id")),
+)
+class InvestorTable(Base):
+    __tablename__ = "investors"
+
+    id = Column(Integer, primary_key=True, index=True)
+    name = Column(String, nullable=False)
+    description = Column(Text, nullable=True)
+    aum = Column(Integer, nullable=False)  # Assets Under Management
+    check_size_lower = Column(Integer, nullable=False)  # Lower bound
+    check_size_upper = Column(Integer, nullable=False)  # Upper bound
+    geography = Column(String, nullable=False)
+    stage_focus = Column(Enum(InvestmentStage), nullable=False)
+    number_of_investments = Column(Integer, default=0)
+    created_at = Column(DateTime, default=datetime.datetime.now(datetime.UTC))
+    updated_at = Column(
+        DateTime,
+        default=datetime.datetime.now(datetime.UTC),
+        onupdate=datetime.datetime.now(datetime.UTC),
+    )
+
+    # Relationship to portfolio companies
+    portfolio_companies = relationship(
+        "CompanyTable",
+        secondary=investor_company_association,
+        back_populates="investors",
+    )
+
+
+class CompanyTable(Base):
+    __tablename__ = "companies"
+
+    id = Column(Integer, primary_key=True, index=True)
+    name = Column(String, nullable=False)
+    industry = Column(String, nullable=False)
+    location = Column(String, nullable=False)
+    founded_year = Column(Integer, nullable=True)
+    website = Column(String, nullable=True)
+    created_at = Column(DateTime, default=datetime.datetime.now(datetime.UTC))
+    updated_at = Column(
+        DateTime,
+        default=datetime.datetime.now(datetime.UTC),
+        onupdate=datetime.datetime.now(datetime.UTC),
+    )
+
+    # Relationship back to investors
+    investors = relationship(
+        "InvestorTable",
+        secondary=investor_company_association,
+        back_populates="portfolio_companies",
+    )
+
+
+class SectorTable(Base):
+    __tablename__ = "sectors"
+
+    id = Column(Integer, primary_key=True, index=True)
+    name = Column(String, unique=True, nullable=False)
+
+
+
+
+class InvestorTeamMember(Base):
+    id = Column(Integer, primary_key=True, index=True)
+    name = Column(String, nullable=False)
+    role = Column(String, nullable=False)
+    email = Column(String, unique=True, nullable=False)
+
+    investor_id = Column(Integer, ForeignKey("investors.id"))
+    investor = relationship("InvestorTable", back_populates="team_members")
diff --git a/app/db/tables.py b/app/db/tables.py
deleted file mode 100644
index 3153b82..0000000
--- a/app/db/tables.py
+++ /dev/null
@@ -1,23 +0,0 @@
-import datetime
-
-from sqlalchemy import Column, DateTime, Integer, String
-
-from app.db.db import Base
-
-
-class InvestorTable(Base):
-    __tablename__ = "investors"
-
-    id = Column(Integer, primary_key=True, index=True)
-    name = Column(String, nullable=False)
-    aum = Column(Integer, nullable=False)
-    check_size = Column(String, nullable=False)
-    sector_focus = Column(String, nullable=False)
-    stage_focus = Column(String, nullable=False)
-    region = Column(String, nullable=False)
-    created_at = Column(DateTime, default=datetime.datetime.now(datetime.UTC))
-    updated_at = Column(
-        DateTime,
-        default=datetime.datetime.now(datetime.UTC),
-        onupdate=datetime.datetime.now(datetime.UTC),
-    )
diff --git a/app/main.py b/app/main.py
index 400730f..88c2143 100644
--- a/app/main.py
+++ b/app/main.py
@@ -1,16 +1,14 @@
 import io
-
+from api import investors
 import pandas as pd
-from app.db.db import db_dependency, init_database
+from db.db import db_dependency
 from fastapi import FastAPI, File, UploadFile
-from app.services.openrouter import InvestorProcessor
-
-from app.pydantic_schemas import QueryRequest, QueryResponseList
-from app.services.langgraph_agent import LangGraphQueryAgent
+from services.openrouter import InvestorProcessor
+from services.querying import QueryProcessor
 
 app = FastAPI()
-
-init_database()
+app.include_router(investors.router)
+# init_database()
 
 
 @app.get("/")
@@ -32,11 +30,11 @@ async def parse_csv(db: db_dependency, file: UploadFile = File(...)):
     return {"results": [r.dict() for r in results]}
 
 
-@app.post("/query", response_model=QueryResponseList)
-async def query_investors(db: db_dependency, request: QueryRequest):
-    agent = LangGraphQueryAgent(sql_session=db)
-    result = agent.run(request.question)
-    return result
+@app.post("/query")
+async def query_investors(db: db_dependency, question: str):
+    processor = QueryProcessor(sql_session=db)
+    results = processor.process_query(question)
+    return {"results": results}
 
 
 if __name__ == "__main__":
diff --git a/app/services/__pycache__/langgraph_agent.cpython-312.pyc b/app/services/__pycache__/langgraph_agent.cpython-312.pyc
index 06c5a6dcea45774bcf19e7b18556d8b25035b0e2..5168061b7e4e3e11bc019fdeaf86c71ac8e8cc2b 100644
GIT binary patch
delta 1539
zcmah}OK2Nc6dnES(QiJ{*!qlqqyNX&$FUqqwqnPzQYXf0oY)E31e&@~NQ%ji)R0Cg
z?Lww4kx;=F7$nhvlvV7y5M`B+MG15>Pa~M;V2CK$6bhYKDrpymj!eAirjKDR_uO~h
zedoUS?jP|taqJhZR)vV;XYJ0XZT&X(yVQk|U(N4&hZ>75@AUmSmUMCQ>D@s1xx5uf
z-_Nj=w<S;SSY`jnvu1Bj-a-vRAHqcU5h<hkX_g<I2cz?lw033En5q}Hh+*2so5H{p
z=1sG}Gz&>zo3OC9bAn7dhzSaWJ}v1W3R;~0E#4&%lHN>mgeRwfsP+N2hUM>87I}9J
zxMMAJ>=4V=<yqc04t(P+bm9=aQ<2xyRdov;5|b*i9{Ky*8`oNB@cj(YTWO)8!})hB
z7hCA?p_i@qx6sI;cjuw3g}T~0!+NTaDo*C)-=R0GVx>No_UhFeJ4WS-B3qLn$bAHQ
z{keXq#(6aj)HJJhLW-#zV|77R6BJa4mgF^NpfU3r2hcbkrZ`RXwN|%wVUHYGHx*2q
z^Sn6(%pr~p^W-QXN4Ksw2A)};TKU90NX&DI583!x_Ur;1JI9gd-%y72GlesoU+}gB
z$l4MdmE@^OKutmuy^$_wN*Uf90p<v29%1RDTT#{!>&T=7gKrfI@<<m$28@UFIIzVz
zD#2670d@TU*tkk$4<cgnWFfiv8E+i{))6kNJId?gKo@UB8%xhFKfTPS7C>r&OPyyY
z7TLumHgSQ|U3{xolb0kNnHIy|nGsC4wv-?F=90i5N|oS2sDbrZAyz!Lk7$)x?j+Rc
z9%R-sdxQlV&1FN$01b4xztj(Hj`F?IJ;=CtCJLA+v^#5fsER|{T27Xd(Bp&tAf)46
zow(CKWg0Ce$ixK&D?(Z~?y$I*MZHCplHOS)UP-gd;tAgTTl~F0JVB~gl8~U*@dC~h
zZa}!VVo#25A7@9VUtVDeH%BbwPPBDYZ>t+Lr*k)T!TOQh3>0^tE~Gbt9OY?JLv=?k
zy^HHgiam>`a=X4z{hYHz_7RL6ETr=%@;4xDDUX%LDp^n6&(WiKkx#E@3Yk5F8!~Qa
zXT+P-li>?Mr@J;*9jk{L4$hf^p$MFsdA{(Xf6KLXecRh$ehj=k2c1)*TBAu+J1{6H
z#4KVtNv7gu@c@ek{^}|b!YDpXCE3;xZNh#S5R@_z+X+)eS@z3bA*=_txE&nM^|dJ@
zPx%1ltH=1@1Q0)JqJxmio>pA<V0CS^(g$!4bougq`CIF&h1Dj(v}vZ|s`;yaKo8~7
zJy#SOnX;v1fd+e7Q_?gIk-Gdb`Ur(KFK-(GwvnbSy|w&gWqTzH89&dA0cNbp%r#tm
zat<WtpxwvYM}d8`X`gM#`S>h|&qCT+o-9o^=|p|`@zqCHp(n(9rhsQ^ACY)NP*3yv
zVW1y|{wVK11^lN(I1oL6BS0VdFT1PPxR3M%_<IUCf)UiB#BKmF+bexFZB_dN&e0KQ
zwZB!!2$g`yl&YQI3?@mu+$&f3;NI*Jrpj!K7#2H~!&CEBeYJ2kyyfN`@tk@Ox7Ww{
z;4}zMxA2oJda^ifdT}}}^1eoP<x2vJ_=n+eROq)Uob3U(pb%}ZQw9$ZF|5KmN)hYK
zNc;+4=YoV*1UpGuq3TY^6f=@yn*N$pE?)`V`sm8FPj248#D!lG#b4<UB$LH}+2MA+
eMpE5AndFotZ|ovH`%aUQN@gVc$bmWNW&Z-{co7W%

delta 1550
zcmai!O=ufO6vuaGKcro)R-fzD$4atRQr5B>sWFyg7sqLx#CCnCsYwW=#3dgMHWV5X
zH$8-)6-ls{6eQz-<r1(;4|PjR8+^&J^w?f@v?@Ceh6<WP4~1?lm9&RKSE7Q_ONM2J
z_n-N_;k`Heo{j%L&j09kvzR@UUmq|3;D5~j2GTD6ZfrlDeUfgbPpVO!Pa62iedU!W
zXPe5j8rOa4CO*9vC;lc&14%XA1b&OgqPas1M9Hs~&HOCP&myPLAwYrbZj1J|5H^EZ
z7|fc%X&9VF&P0p!>46!G0B*{nFyR#5#wfu)ey`B+VlLs~1%n+_GYB}=^5uM0s;!!d
z0!$Q|;MKmCsHaP&T!eD53C8-OSWEBZwsTEz#J-epcHxIYjcbCwE8Frk!BGFaRQsq2
zl)jayD@`!mx5P)OCWyD({>nw|;)8e9^!=e5aXV92KsR6Au@`1Oph{ok7^Y$<nNk%*
zF(xHIO3<koa)oM&&dNF=TTaa3GU)+G516zBY3b2LgU*@s2&6}zk<PXG^8CklOivPe
zk_IW8qyk9=Ih^Z4C9cIyCI*?9!6bC&U>(;fxlI6y?ywle^uUr$W}zo*ki#ZffMnq%
z<a0MeQk{W5Zqq~1Gh~oilgvXh|1w(SZF=xObSVWrDT7R#WEPUym*L3(Nwo>VLB1Hq
zD7JiNUB3IKmBR?q%458za!xz<;QS%xCMfj^BG|R5^3(z8L#(?pqzxg4tB9J2yn!u#
zgGYYB^vlpcDkEQTE3uJ4oToCc%_A|4A`;@Vod8SGe-Io=S%mG-ieq*~M4d_MAi2ph
z(Y?34PeEzfuH((W?CIYUT0kV6$l+Qml}n~Wf(~gn^G*JXyq=x-Zc%qghGR}W)?!_i
zjFzdo3|4Beavi9n$i-EPTCo~6T(O2LRrjmKeVWxiJn%(o#ky~M-0-CjF`ja0)H=RC
zf;hf1qs>%HW3>r`%cy4&#jP!smk#Iv;=Hyq<PW0Y0160*_g1cH*Q!hPMS~wg;V8-#
zzQw;hUOQV)ZlABe{mF$V0P#6HT8Gn)mUdW_eT#Jjb46>jI+NBx`p*u7u>t!BC7dn7
z(PAV04T~XcZ7&<#Bue+YiHP^!ax0^#M_PQq<Ymaq^>b!s4B8(**2a7UtzJbmyS7qZ
zsosKg422|hW_`3Wp-nVMq2&>(SGQI+R-k84ojV9+5bLiT)s7;@SCKTS!6a+JozQj&
zc|)c*1-+?;cdWkpdFj(q3Hc+YUx9w5;h)?cHFJ|NH;H^<)0ct1Ov5*^8#J>MFgt;`
zU}ZsDXmG=|)t%ehw~?4I#bZ!Bc8KAj1afkwGY*|`6p_uyaTqyn%Tc)}C!sU>f4U=b
zVY(YYwi|&#3YN<bu``SX#OkfBI~#X4ONL)Tg7};wSkl4>g53Kh6v4+o>!z#Fdp8Gc
zu(sS}6}w+agFAEEb2|&$3%h3wX++H(Fp>H@W@Zv*CY#I&9h}%KM!ugrZFBdqeSFdG
z3@HUEIm?j%v6d%cQTE#Hmzkc@YaDIH*4n4>qd;yi;ZcxjzZL~#XOaL@o3BUjh258~
zU(a8?oL{|k`Th59-r#NDPlH`;^QY)6equ8w$tmS$?2Qgh;HU6vxPx`q7u+Jirw+01
I+-qz90F=)1;Q#;t

diff --git a/app/services/__pycache__/openrouter.cpython-312.pyc b/app/services/__pycache__/openrouter.cpython-312.pyc
index 46d57bd3895341e45d2e94e48ee40b4cb14b231c..8fcc5d3ee0bea103e577f6578ea577dc81beb601 100644
GIT binary patch
delta 20
acmaFq{L-2GG%qg~0}wnD*}Reao+1EAg$BC-

delta 20
acmaFq{L-2GG%qg~0}$As+qjYYo+1EAMh46P

diff --git a/app/services/__pycache__/querying.cpython-312.pyc b/app/services/__pycache__/querying.cpython-312.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..f3237606f10202ccfa1f8318df891adbe98c8482
GIT binary patch
literal 3970
zcmaJ^&2JmW6`x)1lFJW?5=DJcl1({EM7ky^MRj1iO;E*FVl}e-*~VoLixqc8F1^d8
zXO?zE$bb(%=#T>SNj~ZnxUdfHCH4Q%ix{~ev5-&$>BTpuGSDKYz8P{!Nl~+hnK$pv
zJidAF{oXwOy{9LD;9CfPTlhy1q3`L!{e(7g@LwP{k&0B#Kqi;vI7;&dpXFip7#>r|
z3XCThqUp_g87&$<Q_4z=_8NXOkPR@}X9UeqHpFPj=rMb<y=FKYW^2FEXZB_LIK(4O
zD4uZ*>CS541=m_-Bd|v*e$bV<lp5^piE5GJFI;|iRzqsfcJB+m#ZkjmAhJU|`W&f!
zPmtR0O0vUid@}NuZcB2GbiHirR!K3IVc{uMR$z+UzPWTov6U5tXm>2jxTo8|4Ibjc
zGH|(|*w@QiX?~F&$ze^gH5ta7Eh~AgWG8t?S}Z-#h;3mf<WfsIp~@tp*|uKFlj@YA
zl=8WPqL)&No+0;*6g!<FYsA*fR2f@l*{%+OK)1C^nM_7+>m+G)#qoksz<(>!+whK0
z)fGd_*^X#iW%*w9SAqEFnjKe4<e`R1d=-?%?SdA+ecy<y?HhtrWy}X~(83*lj(W)b
zylvv32*f7RP?l3smiHq+QiCe5c{E`ZB&x@yRs~gfB4$NZ)V!(}xIQWeuJ4JEm43pN
zfa_QNY_AV?1z;Um<B}n#_hv=IYd5f!(+IKfvJ(VX$OIM;Hm9#;w5+bkIYS4p=R9po
z=|ruj>w|5`uT5mPvkW1-oh1krT|ChJ?N8-YUiEAXFWRzoBzLshes`D6Fh*DFsOb8~
z6~eqNx)}%OJ90bhZaMiBa6|5n?p*Be(k}I)O~8qW%j?d0j^@7|bvgEsOKp1(-K+Ck
z>pWQ5<L<F1T&j4=rMk1~+m>839zF$@#Jam&syOLV-C6ak0X3*|+o3}~>fCV+C3}{u
z;|1F;6UcZAK!&Y~t>H8vL@Dc;2eZ}jyk+GLEtA(wy`-l;OrKA!8VV_-C`>R7Z&Gl4
zuvxyRtvM2fvRuK2Bbt_~8BP$eRo1X#SFi@S;Q>k>!!VNq9)_b7K9)foZwq}+7;B_#
zl|aR+g-yk-2Gf~b0l=g{;sKh4vSJsUK^K{_s*xPlSz#eg@X)F{>(+tzAueGMY@IeC
zhbbC~4sf+{6#~`?CB@Vn$<%Cxf;Jh!J}?qPWk*ir`i+I<Ti1VgXW^DSe`8VpV&M;%
z1_9HGfGOr2AH59JFf7rPKd;$x>w1ItDYRr4LhX9s^tMWmYy!rh7Gav_j>xJK9-y?;
zx;yx+?*^0=sAypHWd=WIgsQi&+^&Z?%b@j$LN(B(lO9JPnz4#$;xe6t0xTYn|G1PX
zSf-Y-jLJiWC?!+RTmdgouaL}_7QUCc_QjQqQnIa*{7}c*DhAg;PLwl5!w+<*5gB$V
zD6Q!-#9fvlJ8W654j(NeX;(nQO}>TUF?tt3(NlZT@p?2<i)I>wbB)MMV`TK5UyKGe
zgs-J{AvAJ!`_{`lf4lqFyRU~o0bXctaJms0+Dtr6(COUMbM;7~7D?<z&OTS`>8rK$
z)t%I0EwZ>bFu6CL098X&RiGI{ld~JbYiZ(N@kBj-u@=Ah>e1`?HO6YhPSs<XS}aqK
zU9H8gHu@89`X+Wx-)egJq28v5#um8)B#wciKZOs#atIE)v*X1Y<Z~(4DQ+O6eq|-_
z8z44mse<|VN4PcaWJ$48__}A-IT!qo_;tQW6YbdNmZl=j(_^12d*tbs(sj=^@9Mmb
zlAh`ei&pEYy1CSOk^m7n?xGBc78)VB()9KAyAqNVS!Ovt?JF$@g~RcYoMI@LI3C3?
zSV9BRQ(zgMg@HPbR16#c5*CCOeaOT(PE%2YxmKocj{9I%Cy(XQPq)lbMW4a&7`+J(
z)WfI03!iQb3^%<fGF0!Mto2XU`;)c)<n#A;`)4-1jc9Ch`RVd*bgCXr)S`*q=-Cbb
zUU+Ove;uCLk!J7=s5%l>Y9tHm1`t0NR=X2Gwc}!rrUBJ(>^mu3>s+2=Z(%#B9`Rj`
z>xVV+5IC_q*A3jUcH(DZR~`QS!sR>HZ^duJun_-p{&xKG{L<3GmABM*D1gD1hi&aE
zn>hXwbXhc91Y*eQkyyrh$;RhkGmB@^b11gH_GY)ju(09}U~hHm#5f)+habaE@>>|%
zp`7|IJO!cLlX`Y~r(a76i0Q!aX69+89+|F1rW>QDp7|O>!+_I)!~yaQ1U7u{B;@bi
z$ba42lu+OBMlTDeBf+Hzt&B$kOVPvvOS9+y1@<IV(xzpdTjledegvNkPoK8?r??Vl
zu8da*R7i%erKM}^`DxXTII(dhM_)y?3pASeY5M-K_kcJ;1&VEw#|bde@dE0sdz$O>
z-%y_9zL!Uw?)1QOO1n6j*-#xn=KdqnWza(?Lc#6e-%C#~)d!|(15=Hj;b$LhUD%18
z+Zj#mgwh8d9;&khQRo}l2s6(n`Q_zgkk)?sZpyL~kY#xFD+Z-QGW@z?v?Q2%+VNT|
zMR=odhAuqd09GLIMOd;t!88<qK}>fmduXSE=|Gz*Qi9<L0t=F<Mr!s4UrY?V8Ju{w
zS{r=7DZ&CIWAAN!REwQGpt9x#Nt}J|X(E_jO0ULiGoKw$x)}|N!_ViM2&R`8UM<z;
z77r-hJR^!DTe&8J>ACTaPAyZY0m%`^Yp<0dnCvb%2|UN&`AdN*%AGTvwVbm|vjQ*T
zTAID*tPz0LfUtJKnK)@<#nz#x(lvs;R9Rnl4{gj8oi!s)gtGHkDHqaZtgTdZD5tF7
z=qs<O@LuaVndAzZsSr#%=$4CF2P2d#hT()E`tV0brhS<K|D+s={#k*RO>Z?Z-d3&0
z)ATU&{}QEWVPrA+QTlZ&#-HH3z@wc$xer6r!*Sd`8v7oF_tEG+n%_s~_tEq|n*Ar5
z+eZ@s$$_yyzeiv2ow51d$fu9P4RQ4GAM4_1O&r~ceYh*mJ?7sC19c%*6JlEzUe4|a
fv0dSO!+Yfi4<{~gO%LM2{|SA}jr<$Iz)bod%;e-I

literal 0
HcmV?d00001

diff --git a/app/services/langgraph_agent.py b/app/services/langgraph_agent.py
index 9bf1866..e69de29 100644
--- a/app/services/langgraph_agent.py
+++ b/app/services/langgraph_agent.py
@@ -1,162 +0,0 @@
-from __future__ import annotations
-
-from typing import Any, Dict, List, Optional
-
-import chromadb
-from langgraph.graph import END, START, StateGraph
-from pydantic import BaseModel, Field
-from sqlalchemy import func
-
-from app.db.tables import InvestorTable
-from app.pydantic_schemas import QueryResponse, QueryResponseList
-
-
-class AgentState(BaseModel):
-    question: str
-    sql_results: List[QueryResponse] = Field(default_factory=list)
-    vector_results: List[QueryResponse] = Field(default_factory=list)
-
-
-class LangGraphQueryAgent:
-    """Simple LangGraph agent that queries both SQL and Chroma and merges results."""
-
-    def __init__(
-        self,
-        sql_session: Optional[object] = None,
-        vector_db_client: Optional[object] = None,
-    ) -> None:
-        self.sql_session = sql_session
-
-        # Setup Chroma collection
-        self.vector_db_client = vector_db_client or chromadb.PersistentClient(
-            path="./chroma_db"
-        )
-        self.collection = self.vector_db_client.get_or_create_collection(
-            name="investor_descriptions",
-            metadata={
-                "description": "Investor descriptions and investment thesis focus",
-            },
-        )
-
-        # Build graph
-        graph = StateGraph(AgentState)
-        graph.add_node("sql_search", self._sql_search)
-        graph.add_node("vector_search", self._vector_search)
-        graph.add_node("merge", self._merge)
-
-        # Parallel fan-out: START -> sql_search & vector_search -> merge -> END
-        graph.add_edge(START, "sql_search")
-        graph.add_edge(START, "vector_search")
-        graph.add_edge("sql_search", "merge")
-        graph.add_edge("vector_search", "merge")
-        graph.add_edge("merge", END)
-
-        self.app = graph.compile()
-
-    # Nodes
-    def _sql_search(self, state: AgentState) -> Dict[str, Any]:
-        results: List[QueryResponse] = []
-        if self.sql_session is None:
-            return {"sql_results": results}
-
-        # Simple LIKE-based search across a few fields
-        # Note: SQLite uses case-insensitive LIKE by default for ASCII.
-        q = (
-            self.sql_session.query(InvestorTable)
-            .filter(
-                (func.lower(InvestorTable.name).like(f"%{state.question.lower()}%"))
-                | (
-                    func.lower(InvestorTable.sector_focus).like(
-                        f"%{state.question.lower()}%"
-                    )
-                )
-                | (
-                    func.lower(InvestorTable.stage_focus).like(
-                        f"%{state.question.lower()}%"
-                    )
-                )
-                | (func.lower(InvestorTable.region).like(f"%{state.question.lower()}%"))
-            )
-            .limit(10)
-        )
-
-        for row in q.all():
-            results.append(
-                QueryResponse(
-                    name=row.name,
-                    aum=row.aum,
-                    check_size=row.check_size,
-                    sector_focus=row.sector_focus,
-                    stage_focus=row.stage_focus,
-                    region=row.region,
-                    investment_thesis="",
-                    investor_description="",
-                    reason="Matched SQL fields via LIKE",
-                )
-            )
-
-        return {"sql_results": results}
-
-    def _vector_search(self, state: AgentState) -> Dict[str, Any]:
-        results: List[QueryResponse] = []
-        try:
-            q = self.collection.query(query_texts=[state.question], n_results=10)
-            # q has keys: ids, distances, documents, metadatas
-            docs = q.get("documents") or []
-            metas = q.get("metadatas") or []
-            if docs and metas:
-                for i, md in enumerate(metas[0]):
-                    name = md.get("name", "Unknown")
-                    results.append(
-                        QueryResponse(
-                            name=name,
-                            aum=0,
-                            check_size="",
-                            sector_focus="",
-                            stage_focus="",
-                            region=md.get("headquarters", ""),
-                            investment_thesis="",
-                            investor_description=(docs[0][i] if docs[0] else ""),
-                            reason="Vector similarity in Chroma",
-                        )
-                    )
-        except Exception:
-            # Best-effort; leave vector results empty on failure
-            pass
-
-        return {"vector_results": results}
-
-    def _merge(self, state: AgentState) -> Dict[str, Any]:
-        # Deduplicate by name, prefer SQL fields where available, keep first reason
-        merged: Dict[str, QueryResponse] = {}
-
-        for item in state.vector_results + state.sql_results:
-            if item.name not in merged:
-                merged[item.name] = item
-            else:
-                existing = merged[item.name]
-                merged[item.name] = QueryResponse(
-                    name=existing.name,
-                    aum=existing.aum or item.aum,
-                    check_size=existing.check_size or item.check_size,
-                    sector_focus=existing.sector_focus or item.sector_focus,
-                    stage_focus=existing.stage_focus or item.stage_focus,
-                    region=existing.region or item.region,
-                    investment_thesis=existing.investment_thesis
-                    or item.investment_thesis,
-                    investor_description=existing.investor_description
-                    or item.investor_description,
-                    reason=existing.reason or item.reason,
-                )
-
-        # Store back into sql_results to pass through the END with full state
-        return {
-            "sql_results": list(merged.values()),
-            "vector_results": [],
-        }
-
-    # Public API
-    def run(self, question: str) -> QueryResponseList:
-        state = AgentState(question=question)
-        final_state: AgentState = self.app.invoke(state)
-        return QueryResponseList(responses=final_state.sql_results)
diff --git a/app/services/openrouter.py b/app/services/openrouter.py
index 6a36a61..057a5dc 100644
--- a/app/services/openrouter.py
+++ b/app/services/openrouter.py
@@ -3,13 +3,12 @@ from typing import List, Optional
 
 import chromadb
 import pandas as pd
+from db.models import InvestorTable
 from langchain_core.prompts import PromptTemplate
 from langchain_openai import ChatOpenAI
 from pydantic_schemas import Investor, InvestorList
 from settings import settings
 
-from app.db.tables import InvestorTable
-
 # Add these imports for your databases
 # from sqlalchemy.ext.asyncio import AsyncSession
 # from your_vector_db import VectorDBClient
diff --git a/app/services/querying.py b/app/services/querying.py
index ea80784..e253b66 100644
--- a/app/services/querying.py
+++ b/app/services/querying.py
@@ -1,13 +1,22 @@
 from typing import Optional
 
 import chromadb
+from langchain import hub
+from langchain_community.agent_toolkits import SQLDatabaseToolkit
+from langchain_community.utilities import SQLDatabase
 from langchain_openai import ChatOpenAI
+from langgraph.prebuilt import create_react_agent
 from pydantic_schemas import Investor, InvestorList
 from settings import settings
 
-# Add these imports for your databases
-# from sqlalchemy.ext.asyncio import AsyncSession
-# from your_vector_db import VectorDBClient
+# Connect to SQLite
+
+prompt_template = hub.pull("langchain-ai/sql-agent-system-prompt")
+db = SQLDatabase.from_uri("sqlite:///investors.db")
+system_message = (
+    prompt_template.format(dialect="SQLite", top_k=5)
+    + "\n Get answers from the Sql database and the vector database"
+)
 
 
 class QueryProcessor:
@@ -19,12 +28,16 @@ class QueryProcessor:
         self.llm = ChatOpenAI(
             api_key=settings.OPENROUTER_API_KEY,
             base_url="https://openrouter.ai/api/v1",
-            model="openai/gpt-oss-120b:free",
+            model="google/gemini-2.5-flash-lite",
             temperature=0,
         )
-
-        self.structured_llm = self.llm.with_structured_output(InvestorList)
-        self.sql_session = sql_session
+        self.toolkit = SQLDatabaseToolkit(db=db, llm=self.llm)
+        self.agent = create_react_agent(
+            model=self.llm,
+            tools=self.toolkit.get_tools() + [self.query_vector_database],
+            prompt=system_message,
+            response_format=InvestorList,
+        )
         self.vector_db_client = vector_db_client
 
         self.vector_db_client = chromadb.PersistentClient(path="./chroma_db")
@@ -49,13 +62,22 @@ class QueryProcessor:
         """Query the vector database for investor information."""
         if not self.vector_db_client:
             return None
+        print("VECTOR STORE WAS CALLED")
 
-        # Implement vector database querying logic here
-        results = self.vector_db_client.query(collection=self.collection, query=query)
-        investors = [Investor(**doc.metadata) for doc in results.documents]
-        return InvestorList(investors=investors)
+        # Query the collection directly, not passing collection as parameter
+        results = self.collection.query(
+            query_texts=[query],  # ChromaDB expects a list of query texts
+            n_results=3,  # Specify how many results you want
+        )
+        print(results)
+
+        # ChromaDB returns results in a different structure
+        # results will have 'documents', 'metadatas', 'ids', 'distances'
+        return results
 
     def process_query(self, question: str) -> InvestorList:
         """Process a query using the LLM and return structured investor data."""
-        response = self.structured_llm.predict(question=question)
+        response = self.agent.invoke(
+            {"messages": [("user", question)]},
+        )
         return response
diff --git a/requirements.txt b/requirements.txt
index 6a7dbd3..10ba213 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -8,13 +8,9 @@ chromadb>=0.4.0
 
 # LLM integration
 openai>=1.0.0
-langchain>=0.2.0
-langchain-openai>=0.1.0
-langgraph>=0.2.0
 
 # Environment management
 python-dotenv>=1.0.0
-pydantic-settings>=2.0.0
 
 # Additional dependencies for data processing
 typing-extensions>=4.0.0