VED-011: marriage-election source research

Research checkpoint: 10 October 2026. Research authorized; implementation not started. This packet selects marriage as the first VED-011 purpose family. Construction/Vastu, travel, business, website and Urania remain separate work. Inspected baseline: engine e7107267942d1926da9bcf8ba3d1299b4205ccf9, generated wiki 0c79a578942db2d46e11f203bc6a9c57e730ba27, version 6.9.9. Authority boundary: edition-owned traditional timing rules and their computational implementation. Agreement with a text is not validation of its predictions about a marriage.

1. Findings and admission recommendation

There is substantial usable source material, including in the local library. A complete marriage election is a composition of calendar eligibility, Panchanga restrictions, planetary conditions, wedding Lagna, personal conditions and explicitly scoped exceptions. The existing Muhurta products supply several of these components; neither a generic score nor a list of eleven stars supplies the complete composition.

The recommended first implementation is an ordinary marriage astronomical timing profile owned by the 2004 Avasthi edition of Muhurta Chintamani (MC), with an explicit manifest of required rules. Personal eligibility is a separate, composable layer. An astronomical-only result must not claim to have evaluated the couple. Godhuli needs its own exception profile and cannot be an unconditional override. These are proposed contracts, not admitted profile identifiers or newly available endpoints.

Six consequential findings govern that implementation:

1. MC, Vivaha Vrindavana (VV), Kalaprakasika (KP) and P.V.R. Narasimha Rao (PVR) agree on an eleven-star wedding set which differs from Moira's legacy guidance. The legacy tithi lists also contradict one another.
2. Source differences affect outcomes: lunar dates, weekdays, forbidden quarters, Mars in the eighth, permissible months and Godhuli remedies cannot be collapsed into one supposedly universal policy.
3. MC requires additional calculations, including five/seven-line vedha, Latta, star-history restrictions, Jamitra and distinct Bana formulae. Existing Panchaka, Gandanta and Lagna checks do not cover all of them.
4. Jupiter/Venus asta, balya and vriddha have numerical source support. Surya Siddhanta (SS) IX defines visibility using degrees of time and directional events. Moira's current longitude-orb combustion predicate is a different named convention, not an implementation of that recipe.
5. The ceremony's evaluated instant and the roles assigned to personal conditions must be declared. The source does not authorize silently treating a whole event booking, one partner's birth star or compatibility scoring as the wedding election.
6. A published wedding calendar is only an oracle for its published rule subset. Drik Panchang expressly omits several checks Moira intends to expose. Matching its dates would not establish full-profile correctness.

Decision: proceed to an implementation plan for this finite family using the rule inventory, policy decisions and acceptance gates below. The research packet is complete as a source/contract investigation; VED-011 remains open until the selected engine, public Python and REST contract is implemented and validated. No claim of exhaustive agreement among marriage traditions is made.

2. Research method and source identity

The local corpus was searched by filename and extracted text, followed by inspection of rendered pages at the relevant loci. Image-only books were handled as scans: the Brihat Samhita file has useful extracted text on only one of 1,105 pages, and the 155-page Raman Manual has no extracted text. An empty search result in those files is not evidence that a rule is absent.

Seven local-library PDFs and five additional source PDFs were fingerprinted. Relevant passages were collated; this was not a cover-to-cover review of all twelve volumes. The four electional scans reused from the 8 October research were identified again by their title pages and hashes; online catalogues, digital texts, the PVR author site and the published Drik calculation policy were checked during this pass. A fresh SS scan was downloaded and its heliacal chapter inspected visually.

Private working evidence is under C:\dev\outputs\ved011-marriage-research-2026-10-10: source_manifest.json, page-indexed extraction JSON, rendered page images, and the research probe and its receipt. PDF page numbers in this document are one-based; printed page numbers and chapter/verse identifiers are separately stated. Source PDFs and whole-book extracted text are not added to the repository.

ID	Source, edition and access	Inspected loci and evidential use
MC	Rama Daivajna, Muhurta Chintamani, Ramlal Avasthi commentary, Chandika Prasad Avasthi editor, Tejkumar Book Depot, Lucknow, tenth edition 2004; 207-page scan. Archive catalogue; digital Sanskrit witness.	Samskara 26–28, printed 85–86/PDF 94–95; Vivaha 12–16, printed 104–106/PDF 113–115; wedding-star and vedha rules 55–58, printed 127–130/PDF 136–139; detailed restrictions 59–74, printed 131–140/PDF 140–149; Lagna/exception rules 75–101, printed 140–153/PDF 149–162. Main proposed authority; verse, commentary and implementation convention must remain distinguishable.
VV	Keshava, Vivaha Vrindavana, Sitarama Jha's Vasantalakshmi Sanskrit/Hindi commentary, Master Khelari Lal & Sons, Benares; cover says first edition Samvat 1991; Archive catalogue dates it 1935; 138 pages. Catalogue.	Nakshatra-shuddhi 1.3–4, printed 2–3/PDF 15–16, including explicit Pushya/Purva Phalguni objection; Chandra-bala discussion printed 45/PDF 58, verses 7–8 argues for the groom's Moon as well. Godhuli locus is also owned by the earlier named-window research; its numerical interval is not attributed to BS.
KP	N. P. Subramania Iyer, Kalaprakasika, Sanskrit/English, Asian Educational Services 1982 reprint; introduction dated January 1917; 490 pages. Source PDF.	Chapter XIV, printed 79–87/PDF 116–124. Independent marriage rule collection with explicit best/middling/avoid distinctions and conflicting placements. Online PDF retrieval timed out in this pass; the previously downloaded, fingerprinted scan was inspected locally.
R	B. V. Raman, Muhurtha, 110-page reflowed transcription, underlying printing not established. Available copy.	Marriage passage PDF 64–65. Useful modern witness and variant lead; not sufficient to lock an edition-specific production profile until a printed edition is collated.
BS	Brihat Samhita, V. Subrahmanya Sastri and M. Ramakrishna Bhat, V. B. Soobbiah & Sons, 1946; local Classics books/brihatsamhita.pdf, 1,105 pages. Independent 1884 Iyer translation.	Chapter 103, printed 764–768/PDF 781–785: wedding house placements and Godhuli. The 1946 chapter heading explicitly says “of Vindhyavasin”. Cite the chapter as presented in this edition, rather than making unqualified authorship claims for every verse. PDF offsets elsewhere in the volume differ.
PVR	P. V. R. Narasimha Rao, Vedic Astrology: An Integrated Approach, local Good books/vedic_astro_textbook.pdf, 515 pages. Author's book page identifies publication in 2000; author-hosted PDF.	Chapter 36, printed 472/PDF 484 on the ritual anchor; Table 79, printed 474/PDF 486 on wedding limbs; printed 475/PDF 487 qualifies the empty-seventh guideline. Modern author's stated practice, not automatically a classical verse.
PS	P. S. Sastri, Text Book of Scientific Hindu Astrology, Ranjan Publications; exact printing not established; local Books by Authors/BV Raman/Scientific-hindu-astrology.pdf, 521 pages.	Printed 100/PDF 95 and adjoining discussion: general Muhurta guidance. The folder's Raman label is not the book's authorship. General rules are not automatically wedding rules.
MANUAL	B. V. Raman, A Manual of Hindu Astrology, local Books by Authors/BV Raman/A Manual of Hindu Astrology by BV Raman.pdf, 155-page image-only scan.	Front matter/contents inspected; not used as a wedding-rule authority and not declared exhaustively searched.
UK	Local Classics books/Uttara-kalamrita-kalidas.pdf, 166 pages.	Marriage-related passages located around PDF 123–128 concern kinship/social/ritual context; not used to manufacture numerical election rules. No whole-book absence claim.
BPHS1/2	Santhanam BPHS local Classics books/BPHS-Santhanam-Vol-1.pdf and BPHS-Santhanam-Vol-2.pdf, 482/552 pages.	Targeted text search and existing owned dignity/Navamsa/strength standards. Natal marriage indications or Dasha timing do not by themselves prescribe a marriage election. No claim of a complete wedding policy derived from these searches.
SS	Ebenezer Burgess, Translation of the Surya-Siddhanta, American Oriental Society, 1860; 393-page scan. Source PDF.	Chapter IX.1–11, printed 221–224/PDF 237–240. Primary astronomical definition and thresholds for heliacal disappearance/reappearance. Chapter VIII commentary printed 210/PDF 226 also discusses the unequal Abhijit insertion.

Two collation hazards were caught by checking page images. PVR Table 79's row for placing idols has a different tithi/weekday list immediately below the wedding row; extracted row order must not transfer it to weddings. In MC's seven-line vedha, the actual pairs include Dhanishtha–Vishakha and Chitra–Purva Bhadrapada. Merely checking that a table contains 28 unique stars would not catch swapping those partners.

3. The shared wedding-star set and the legacy defect

MC Vivaha 55, VV 1.3, KP XIV and PVR Table 79 support these eleven:

Star	One-based ordinary 27-star ordinal
Rohini	4
Mrigashira	5
Magha	10
Uttara Phalguni	12
Hasta	13
Swati	15
Anuradha	17
Mula	19
Uttara Ashadha	21
Uttara Bhadrapada	26
Revati	27

This is a necessary list condition in those passages, not sufficient election eligibility. MC also requires freedom from vedha; other sources attach different quarter exclusions. VV 1.4 explicitly argues against treating Pushya or Purva Phalguni as generally acceptable wedding stars.

At the inspected baseline, legacy guidance contains:

• good_nakshatras: Pushya and Shravana instead of Magha and Mula;
• avoid_nakshatras: Mula, despite its inclusion in the collated eleven;
• good_tithis: [1,2,3,4,6,7,9,10,11,12,13];
• preferred_tithis: [2,3,5,7,10,11,13];
• avoid_tithis: [0,3,8,13,29], with a zero-based comment.

Raw values 3 and 13 appear in both good and avoid irrespective of how an index convention is subsequently interpreted. “Preferred” contains 5, which is absent from “good”. The source comment does not resolve these conflicts. get_muhurta_guidance_for_activity returns this dictionary directly.

This is a real legacy data/documentation defect. Existing REST provenance explicitly reports activity_guidance="not_admitted", so it is not evidence that an already-admitted complete wedding endpoint is returning these rules. The implementation plan should replace or explicitly deprecate this guidance under a documented compatibility decision, with a fixture for the exact eleven-star set and a single explicit tithi representation. This research pass records the defect and does not silently change the legacy API.

4. Disagreements that must survive composition

Axis	Evidence	Required treatment
Tithi	MC 55 excludes Rikta 4/9/14 and Amavasya at this locus. KP distinguishes best 2/3/5/7/10/11/13 from middling dates, and separately rejects the dark half after its eighth. PVR wedding row includes 12 and 15. Raman's transcription rejects 6/8/12 and dark 11 through Amavasya.	Separate paksha-aware profiles; do not present the MC locus alone as the exhaustive tithi doctrine or import PVR's 12 into Raman. Resolve broader prohibitions before positive preferences within each profile.
Weekday	MC Avasthi's explanation of 55 and PVR list Monday, Wednesday, Thursday, Friday. KP rejects Sunday/Saturday/Tuesday. Raman grades Sunday/Saturday middling.	Best/middling/forbidden are distinct; no universal four-day filter claimed for every source.
Star quarters	Raman excludes first Magha/Mula quarters and last Revati quarter. MC's selected Gandanta calculation has its own precise intervals.	A whole pada is not the same object as a ghati-based junction restriction. Preserve both only when the selected profile requires both.
Solar/lunar calendar	MC Vivaha 13 names Sun in Aries, Taurus, Gemini, Scorpio, Capricorn, Aquarius and specific lunar-month qualifications, including the opening Ashadha portion through Shukla Dashami with Gemini Sun. MC Samskara 26 has a general Dakshinayana restriction. KP calls Dakshinayana middling.	MC's specific Scorpio permission and its general restriction require a disclosed precedence rule. Six solar signs alone do not implement the lunar clauses. Do not apply a modern Chaturmas switch as if it were an identical verse.
Mars in eighth	KP chapter XIV's eighth-house result is favorable to Mars. BS 103.8, Raman and the selected MC restrictions treat it adversely, subject to their own exceptions.	Preserve the disagreement. Do not rewrite the shared planet-house detector or import a cancellation across sources.
Seventh house	MC includes Jamitra from Lagna and Moon and a more precise Navamsa treatment. PVR gives an empty-seventh guideline, then qualifies it for strong benefic configurations.	Existing Lagna seventh occupancy is only part of MC Jamitra. PVR's qualitative qualification does not define an arbitrary Shadbala threshold.
Karanas	PS's general guidance rejects the final five including Kinstughna. Drik's marriage policy rejects Vishti, Shakuni, Chatushpada and Nagava and explicitly allows Kinstughna. Existing Bhadra checks address Vishti.	A Vishti-only result must not claim full marriage Karana eligibility; a modern site's choice is not a silently universal classical rule.
Birth context	MC 12 assigns Jupiter to the bride, Sun to the groom and Moon to both. VV argues for the groom's Moon as well. MC 14 and KP differ over first-born birth-period restrictions.	Require explicit roles and needed birth/context data. Missing partner evidence is unavailable, not favorable. Family status cannot be inferred from ephemerides.
Godhuli	BS 103.13 gives a twilight remedy without a minute formula. MC 99–101 combines broad relief, seasonal descriptions and retained planetary/weekday qualifications. VV owns the existing half-ghati each side of half-set geometry.	Name both temporal and exception authorities. An existing Godhuli window does not automatically waive every wedding rule.
Regional and extraordinary context	MC 69 and commentary distinguish regional handling; MC Samskara's Ketu/utpata context includes extraordinary omens.	Do not infer a school from longitude or reinterpret every historical “Ketu” occurrence as the calculated lunar node. Explicitly exclude non-ephemeris omen judgments from the first astronomical profile.

Historical social classifications, age prescriptions, kinship eligibility, family ritual spacing and predicted personal misfortunes are not proposed as numerical marriage-eligibility outputs. Their presence in a chapter is recorded; the first product's object is the timing calculation. Compatibility matching remains candidate C-01, a different object from natal-to-election Tara/Chandra strength.

5. MC rule inventory and engine ownership

“New” below means not supplied by the presently admitted marriage composition, not that no arithmetic helper anywhere in Moira could be reused. The final implementation manifest must identify every required rule and every explicit exclusion. Passing a subset must not be relabelled “complete MC Vivaha”.

MC locus	Governing object	Existing component / required work
Samskara 26–28	Guru/Shukra asta, balya, vriddha; variants and event context	New marriage availability layer over a named astronomical event definition. See section 7.
Vivaha 12	Bride Jupiter, groom Sun, both Moons relative to birth context	Reuse canonical sidereal positions and inclusive sign arithmetic; add source-specific role rules and personal completeness. Do not reuse a Moon-only Gochar verdict for the Sun/Jupiter roles.
13	Solar signs and lunar-month qualifications	Reuse lunar-month owner; add marriage predicates and declared month/ayanamsa/precedence policy. Exceptional unavailable calendar evidence must propagate.
14–16	First-born/birth-period and family ceremony context	Optional separately declared personal/context extension; not inferred from a chart. Keep source-specific first-born differences.
21 onward, especially 24	Couple matching, including Tara Kuta	Separate from election Tara Bala; excluded from ordinary astronomical-only assessment.
55	Eleven permitted wedding stars, tithi and related qualifications	New typed purpose rule; reuse Panchanga identities, not the contradictory legacy dictionary.
56	Pancha-shalaka vedha and opposing padas	New edition-owned star-pair/pada relation; unequal Abhijit partition.
57–58	Sapta-shalaka vedha, planet nature, occupied/pierced/passed stars and Moon clearance	New relation plus event-history evidence. Single-epoch longitudes cannot settle the history clauses.
59	Planet-specific forward/backward Latta	New counted-star rule with explicit origin inclusion and 27/28 counting selection from its own locus; do not silently reuse the seven-line grid.
60	Pata and quoted alternative prescriptions	New named rule variants; ending a yoga and its affected star require event context.
61	Krantisamya sign-pair prescription	New source relation. A similar English name does not authorize replacing it with physical declination equality.
62	Ekargala under selected yogas and odd star separation	New rule using its expressly Abhijit-inclusive count and source yoga list.
63–64	Upagraha and pada qualification; adverse eighth of daylight	New source rules; full daylight parent and source weekday convention needed. “Upagraha” here must not alias an unrelated calculated point by name alone.
65	Kulika by day/night fifteenths	Reuse solar-day parents; add source segment indices and Saturday's additional qualification. Distinct from Gulika upagraha longitude.
66	Dagdha tithis by solar sign	New small, source-fixtured mapping.
67–68	Jamitra from Lagna/Moon; 55th Navamsa; specific relief	VED-010 covers selected Lagna occupancy, not all these source conditions. Preserve the precise and broader formulations as named rules.
69–71	Regional treatment, Sun/Moon nakshatra combinations and scoped relief	New source-specific composition. Source relief must target the named detected restrictions.
72–74	Two Bana calculations and their contextual applicability	New rules; neither is the existing Panchaka sum. Day/night and purpose applicability remain part of each finding.
75–78, 84–88, 92	Aspects, Navamsa, placements, selected exceptions, source placement score	Reuse VED-010 Lagna owner and dated owner. It already exposes four/five-Navamsa choice and optional source exceptions. Its placement score is not the aggregate wedding verdict.
79–83	Additional Lagna qualities and restrictions	New admission and fixtures required before a broad chapter-completion claim. The passage is located; source availability is not the obstacle.
88–91	Kartari and wider cancellation prescriptions	Existing Lagna catalogue explicitly excludes Kartari and broad 89–91 relief. Add them as individually sourced rules, or name their exclusion and limit the profile claim.
93–95	Strength/context and special marriage-form qualifications	Do not convert prose into an unexplained numeric cutoff or universal exception. First profile should state the selected ordinary-ceremony boundary.
96–98	Preparatory acts/mandapa	Separate event targets; not automatically evaluated at the wedding anchor.
99–101	Godhuli timing and exceptions	Reuse named-window geometry where source-compatible; build a separate exception policy with retained restrictions.

The related Panchanga Shuddhi, named-window and Lagna research remain the owners of their admitted components. VED-011 should compose their canonical results rather than rebuilding them in a server serializer.

6. Concrete arithmetic fixtures recovered from the source

These transcriptions are research fixtures for a forthcoming implementation; their arithmetic checks do not certify a production evaluator.

6.1 Abhijit and seven-line vedha

MC 55's commentary defines Abhijit from the final quarter of Uttara Ashadha through the first fifteenth of Shravana. Under the ordinary equal-angle 27-star sidereal partition, this yields 276°40′ through 280°53′20″. This conversion is arithmetic derived from the specified portions, not a claim that Abhijit is an equal twenty-eighth star. Boundary ownership must be declared (recommended half-open intervals). Count-grid identity and physical longitude membership must both be visible in evidence.

MC 57, printed 129/PDF 138, gives these fourteen reciprocal pairs:

First star	Pierced counterpart
Jyeshtha	Pushya
Shatabhisha	Swati
Purva Ashadha	Ardra
Revati	Uttara Phalguni
Dhanishtha	Vishakha
Uttara Ashadha	Mrigashira
Ashwini	Purva Phalguni
Ashlesha	Anuradha
Hasta	Uttara Bhadrapada
Rohini	Abhijit
Mula	Punarvasu
Chitra	Purva Bhadrapada
Bharani	Magha
Krittika	Shravana

MC 56's five-line rule has its own pairs and opposing pada qualification: Rohini–Abhijit, Bharani–Anuradha, Uttara Ashadha–Mrigashira, Shravana–Magha, Hasta–Uttara Bhadrapada, Swati–Shatabhisha, Mula–Punarvasu, Uttara Phalguni–Revati; pada 1 opposes 4 and 2 opposes 3. Do not merge this table with the seven-line table into an unlabeled union. MC 57–58 distinguishes cruel/benefic effects and time-history conditions; the named planet-nature profile must be recorded rather than assumed.

6.2 Latta and Bana are not generic Panchaka

MC 59 and the selected Avasthi commentary prescribe forward counts for Sun 12, Saturn 8, Jupiter 6, Mars 3, and Rahu 9; backward counts for Mercury 7, full Moon 22 and Venus 5. The Rahu direction follows the commentary's explicit qualification and is not a generic reversal for retrograde motion. Every fixture must state whether the occupied star counts as one and which grid is used. Changing direction merely because a planet is retrograde requires its own source rule; it is not inferred from the word “backward”.

MC 72's southern Bana rule combines elapsed tithi and Lagna, modulo nine: remainders 8 illness, 2 fire, 4 royal, 6 theft, 1 death (printed 139, PDF 148). This corrects the initial research transcription. It differs from the existing Panchaka input sum of running tithi, weekday, nakshatra and Lagna. Elapsed versus running tithi changes boundary outcomes and must be transcribed into a worked fixture before admission.

MC 73 uses elapsed whole degrees within the Sun's sign, separately adding 6/3/1/8/4 and checking remainder 5 modulo nine for those five Bana types. MC 74 qualifies the applicability by night/day, weekday and activity; these are not five unconditional all-day wedding vetoes. Its commentary identifies Saturday royal, Wednesday death, Tuesday fire and theft, and Sunday illness associations (printed 140, PDF 149) and distinguishes purpose-specific prohibitions. These correct the initial weekday transcription.

6.3 Source-sized temporal partitions

MC 64's adverse daylight eighths by Sunday through Saturday are 4, 7, 2, 5, 8, 3, 6. MC 65's daylight Kulika fifteenths are 14, 12, 10, 8, 6, 4, 2, with preceding night fifteenths 13, 11, 9, 7, 5, 3, 1 and an additional Saturday final-night qualification. The relevant full day/night must be determined before clipping to a request.

MC 66's Dagdha pairs are Sagittarius/Pisces → 2, Taurus/Aquarius → 4, Cancer/Aries → 6, Virgo/Gemini → 8, Leo/Scorpio → 10, Capricorn/Libra → 12. These are tithi identities, not zero-based array indices.

The private research probe checks the eleven-star transcription against the legacy AST, the seven-line pair coverage/reciprocity, these small mappings and the rational Abhijit conversion. Visual collation is still the authority for the table: uniqueness alone cannot establish the correct pairs.

7. Jupiter/Venus availability: a source-backed computational route

7.1 Event definition

SS IX.1–5, Burgess printed 221–222/PDF 237–238, distinguishes heliacal disappearance/reappearance from ordinary daily horizon crossings. It uses the ascensional interval at sunset/sunrise, with the planet's apparent-place correction, converted to degrees of time (kalabhaga). The commentary explicitly distinguishes this from longitude separation.

SS IX.6–9, printed 223/PDF 239, supplies:

Planet/event branch	Threshold in degrees of time
Jupiter	11
Venus setting in the west or rising in the east	8
Venus setting in the east or rising in the west	10

The chapter also gives Mars 17, Saturn 15 and Mercury 12/14 with its branch qualification. Those are corroborating context, not a request to extend the first marriage availability feature to every planet. IX.10–11 describes finding the time of the event from the interval's rate of change.

This answers the source-availability question: the thresholds and underlying astronomical quantity are available. Implementation still needs a declared adapter. Reproducing the historical SS ephemeris and using modern Moira positions in source-derived rising/setting geometry are distinct policies. The recommended first modern adapter must disclose that derivation, its coordinate/horizon conventions, event equality handling and polar limits. Independent worked examples must verify the translation from SS's units.

Moira's avastha combustion predicate currently uses Sun-relative longitude orbs; Venus has a fixed 10° direct-orb convention. Numerically sharing Jupiter's number 11 does not make the two quantities identical. The general heliacal owner has astronomical machinery and modern physical visibility models, but those models must not be relabelled SS IX either. An adapter must name the model it actually executes and preserve the event evidence.

7.2 Waiting periods

MC Samskara 27, printed 86/PDF 95, distinguishes:

Body and direction	Balya after appearance	Vriddha before disappearance
Venus in eastern sky	3 days	15 days
Venus in western sky	10 days	5 days
Jupiter	15 days	15 days

Samskara 28 reports further 10-day, 7-day and urgent 3-day alternatives. They are variants, not permission for the implementation to choose the shortest interval automatically. KP XIV instead specifies a week after Venus's rise, eight days after Jupiter's rise and fifteen before Jupiter's setting; its “just prior” Venus wording does not supply a numeric interval. Do not invent a duration and present it as KP's number.

Implementation must distinguish the actual invisible interval from the before/after buffer, retain event direction, and resolve adjacent events outside the requested date range. A buffer may affect the requested day even when the heliacal event lies outside it. Fixed elapsed days versus local sunrise-counted days is an additional explicit computational choice; the passage does not independently settle that modern boundary convention. The first profile should name one convention and test both edges, rather than hiding it inside a date comparison.

8. Ceremony, partners and exception policy

PVR printed 472 acknowledges ambiguity about the wedding's exact moment and proposes the tying of mangalasutra in the described practice. This is the modern author's proposal. Moira should accept an explicit ritual anchor label and evaluate that instant or the candidate range for that anchor. A start time, contractual booking interval, vows and preparatory rituals are different targets. Requiring an entire ceremony to fit inside an eligible interval should be a separately explicit query mode.

The personal layer should accept two participant records plus explicit traditional role assignments where the selected rule needs bride/groom roles. No role should be inferred from a name. The first implementation can evaluate ordinary astronomical timing without birth data, but must return personal_assessment_unavailable rather than an overall personal pass. If personal assessment is requested, neither missing participant nor missing birth-month/first-born context may silently disappear from coverage.

MC Vivaha 12's bride-Jupiter, groom-Sun and both-Moon conditions must remain individually inspectable. Election Tara Bala and couple-to-couple Tara Kuta must have different input/result identities. The shared Tara profile's existing purpose-dependent Janma uncertainty is not resolved by simply renaming the current result “marriage”.

For cancellation, preserve:

• the detected restriction and the source evidence that detected it;
• the named exception, its exact target rule IDs and prerequisites;
• whether every prerequisite was evaluated and satisfied;
• the remaining effective restriction and any unavailable prerequisites.

A named auspicious yoga, Godhuli presence or a high placement/Shadbala score must not erase all raw findings. PVR's benefic-seventh qualification, MC's specific placement remedies, and BS's Godhuli verse belong to different authorities. Combining a VV temporal window with an MC exception is an explicit derived composition, not a falsely single-source profile.

9. Proposed engine, public Python and REST contract

The engine should own three objects: a discoverable profile catalogue, an instant assessment, and a bounded dated interval assessment. The public facade and typed REST transport should preserve those same objects. Exact function/route names belong in the implementation plan; none are admitted by this research document.

Each result needs the profile/version, source loci, ephemeris/frame/time receipt, ritual anchor, participant coverage and per-rule findings. Aggregate decisions should distinguish passes selected profile, restricted, and indeterminate. Coverage completeness is a separate field: a known restriction can establish “restricted” even when another required condition is unavailable, but incomplete evidence cannot establish a pass. Excluded optional layers must remain visible without being called evaluated. No “marriage suitability percentage” is justified by this packet.

The first profile should lock these axes instead of accepting an arbitrary bag of favorable switches:

Axis	Proposed decision or explicit admission question
Governing edition	MC Avasthi 2004 for the ordinary marriage timing profile. Other books remain separately named variants.
Calendar precedence	Prefer the marriage-specific Vivaha 13 prescription over the general Samskara prohibition where they conflict; label this as the composition decision, not a new textual quotation. A conservative intersection would be a different profile.
Astronomical conventions	Explicit ayanamsa, node model, coordinate frame, solar-event horizon and resource provenance, consistent across component receipts.
Star grids	Rule-owned ordinary 27-star versus unequal Abhijit-inclusive 28-star identity; never a global switch applied indiscriminately.
Weekday/time	Sunrise-owned traditional weekday for rules that require it; preserve UTC/JD weekday in existing Shadbala evidence where that owner requires it. A civil midnight must not silently change the traditional day.
Asta and buffers	Source-derived SS event adapter and the selected MC 27 buffer variant, with an explicit day-count convention. Fixed-orb compatibility must be separately named if offered.
Lagna/Navamsa	Reuse the existing MC source profile and its explicit four/five-Navamsa choice; extend missing required wedding restrictions rather than implying they already exist.
Personal assessment	Explicitly requested layer with both participants and role/context completeness. No matching-score substitution.
Remedies	Individually opt-in or fixed by the selected named profile; finite target lists, no score-based global cancellation.
Godhuli	Separate temporal/exception profile; keep the ordinary-profile result available for comparison.
Context exclusions	Explicit ordinary astronomical timing scope; matching, social eligibility, non-ephemeris omens and other ritual events outside this first contract.

The dated evaluator must split on every transition that can change a selected rule: Panchanga limbs, star/pada boundaries, all participating planetary star/vedha changes, Sun ingress, Lagna/D9/aspect transitions, solar day/night boundaries, lunar-month boundaries, heliacal events and their buffer edges. History-dependent conditions need a bounded lookback or an unavailable result. Compute complete parents before request clipping. Midnight, DST changes and longitude wrap must not corrupt event ordering.

The existing Muhurta search explicitly owns a bounded sampled-score product. It does not already supply this continuous intersection, sunrise-owned calendar or complete marriage policy. A sampled mode could remain available with its sample provenance, but must not export unverified spans between samples as exact eligible intervals.

Set work limits after profiling the actual selected transition inventory. Report horizon unavailability, unsupported calendar branch, missing kernel, budget exhaustion and no eligible interval as different outcomes. Reuse the serving ephemeris reader; retain borrowed-reader ownership and avoid silent resource fallback. REST must strictly validate finite numbers, booleans-as-numbers, unknown profiles/fields, contradictory receipts and unsupported partial contexts. Server code must not repair, score or override canonical engine findings.

10. Validation and independent comparison

Acceptance requires source fixtures plus computational and transport checks; neither internal consistency nor a commercial calendar is sufficient alone.

Test family	Required evidence
Source transcription	Exact eleven stars; paksha-aware tithi identities; all fourteen seven-line pairs including Dhanishtha–Vishakha and Chitra–Purva Bhadrapada; independent five-line table; every numerical mapping traceable to page/verse.
Boundary partitions	Rational Abhijit endpoints, both adjacent stars/padas, 0°/360°, all applicable half-open edges; no equal-28 partition.
Distinct formulae	MC Latta direction/count fixtures; MC 72 elapsed-tithi Bana versus running-tithi Panchaka; MC 73 solar-degree Bana; daylight eighths versus fifteenths; opposite test outcomes where formula substitution would look plausible.
Calendar precedence	Scorpio under the chosen MC-specific precedence; Gemini/Ashadha Dashami boundary; ingress inside a lunar date; adhika/kshaya and unsupported month cases; no six-sign-only false completeness.
Asta model	Source-owned worked time-degree examples, both Venus branches and Jupiter; equality/just-inside/just-outside tests; geographic latitude effects; explicit divergence from fixed longitude-orb combustion; event direction and buffer crossings.
Personal completeness	Both participants; swapped roles; missing Moon/Sun/Jupiter or birth context; Tara Bala versus Tara Kuta; unknown prerequisite must never produce a personal pass.
Exceptions	Raw findings retained, exact target mapping, missing prerequisite, competing restrictions, Godhuli with a retained prohibition; a high generic score cannot neutralize them.
Source disagreements	The same synthetic case may legitimately pass one named policy and fail another: Sunday, tithi 12, Mars eighth, Pushya, and quarter/Gandanta cases must have explicit expected policy-specific outcomes.
Dated composition	Every governing event class, full parents before clipping, negative timezones, DST, before sunrise, polar conditions and event-history limits; dense sampling is a useful additional witness, not the definition of exact intervals.
Engine/HTTP/resources	Direct/facade/HTTP equality of all findings and receipts; strict hostile input tests, reproducible bounds and serving-reader proof; unavailable versus empty interval distinguished.

Drik's published wedding calculation policy is useful specifically because it states exclusions: its published wedding dates do not incorporate tithi, weekday or Lagna selection, and it excludes Simhastha Guru and Holashtak checks from that calculation. It also documents its own solar/month and Karana choices. A future external fixture must carry location, timezone, source date/version and the exact shared rule subset. Differences outside that subset are not automatically Moira defects.

No full paired external marriage-election oracle was established in this research pass. That does not block source-owned arithmetic fixtures and independent astronomical comparisons. It does prevent claiming that a whole profile has already been externally certified. Before implementation closure, assemble worked cases that expose both the input positions and the source decisions, with disagreements classified rather than edited away.

11. Next implementation package and closure boundary

The next plan should cover these coherent units, with one final end-to-end acceptance gate:

1. Profile and legacy repair: exact source manifest, rule coverage, strict typed catalogue, one tithi representation and an explicit migration for the legacy guidance defect.
2. Missing wedding calculations: unequal star grids, vedha/Latta and associated doshas, distinct Bana/Kulika/Dagdha rules, source calendar predicates and the remaining selected Lagna restrictions.
3. Jupiter/Venus availability: declared SS-derived event geometry, directional MC buffers and independent astronomical fixtures. This is a real source-backed dependency, not “classical evidence unavailable”.
4. Composition and personal scope: ordinary astronomical assessment, optional complete personal layer and finite cancellation targets; separately named Godhuli behavior.
5. Bounded dated service and parity: transition-owned intervals, resource/work limits, public/facade/REST identity, adversarial acceptance, owning standard and validation receipt.

Before code changes, the plan must turn every “new” or partially covered row in section 5 into either a concrete required rule or an explicit exclusion that narrows the product's name and completeness claim. No required but unimplemented rule may be silently converted into an optional omission. Literal historical astronomy versus modern source-derived geometry, time buffer counting and qualitative cancellation prerequisites require named decisions and worked fixtures; the relevant source passages are available.

This packet does not require a mechanical twelve-phase programme. It does require a complete selected contract, including REST and evidence, before calling that selected marriage-election profile done. VED-011 as a wider purpose-election family can retain other named activities as separate work.

12. Research verification receipt

This pass changes only this canonical research document and the remaining work register. Runtime, tests, protected validation fixtures, dependencies, version and generated publication outputs are unchanged. The source evidence was inspected with the project's Python 3.14.3 environment and PDF rendering; no production election evaluator was executed because none was added.

The private probe passed: twelve source files were rehashed; the legacy AST contradictions, source-table consistency and rational Abhijit boundaries were verified; 111 local document links resolved. An in-memory preview through the owning Git-wiki generator passed for both changed pages, explicitly including this new untracked page. The preview verifies link rewriting without changing the generated mirror.

Executed check	Result
Private research_probe.py	Pass; receipt in research_probe_receipt.json
scripts/check_doc_consistency.py	Pass
scripts/check_release_identity.py	Pass for 6.9.9
scripts/generate_hellenistic_inventory.py --check	Pass; inventories current
scripts/sync_rest_api_reference.py --check	Pass; route inventory current
git diff --check and explicit new-page whitespace check	Pass

All Python checks used the project .venv, with UTF-8 output and MOIRA_NO_DOWNLOAD=1. Runtime regression suites were not rerun for this documentation-only research. This receipt does not inherit the previous repair package's test count as a new run. Actual generated-wiki synchronization and its check belong to the later publication preparation; only the in-memory preview was used here. This research does not stage, commit or push the packet.

Source fingerprints

The following SHA-256 values identify the inspected PDFs, not a claim that every page was read. Paths for local-library witnesses are given in section 2; remaining cached paths are in the private source manifest.

ID	SHA-256
MC	9e931eb7fdc1958516424590a168680be60753f4cfccea8f327cae973dfe4777
KP	d082d5b3b29cb6c054f3cfe8e5ff50dfa6a66220128b52afa485c5a91c10c074
VV	793cc3c454a0c383bb3784886cd19e702d310b67aa39a597748b148da02cdfc0
R	8ab2c1fc0a4ca0e0f7145434b4c02f8efa97ce48d27fc03732cd632d6a3f9256
BS	30345203af3be094776ef67f7a9ad2ae8da8392513c72ac7d24a791eb7b19302
PS	e82c5936775a5d3577433bb88271667ef18580bf449505b5bffa6fd12fa54bdf
MANUAL	439f877fd4a52f309f67e9e427a912daff7241cfdfb180605e6e6eba9af5b60a
PVR	2971851a3cda30929cf5b1ec56b88902bd4a141c8d5479599573ddf1c5f78242
UK	38dac73790e3c21d289bddb4cefcf558e3944b5b0e691bfad037de8c06979386
BPHS1	13609f1af89541e070c798b8c70c6bfd74a8b9a71a71388913bf898e9f9ddf3f
BPHS2	ad7172b615568eb812961c71cff352f354c5676b955bc34cfcfdcf431404cd45
SS	2eb5c967bb5363e4e5d3e8c9a0e51fc92855af3b6d023166dcc0526ec25b930b