write manuscript structural.py

import json
import re
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]

modules = [
    {
        "id": "ordered_measurement_esau",
        "title": "Recovering a complete list from ordered measurements",
        "text": """Esau’s gift shows how a narrative list enters the same grammar as a chronological path. The source supplies nine printed counts, their species and sex labels, and the clean/unclean classification. It also supplies distinct orders: the male-first clean walk, the female register in drove order, and the unclean framing walk. Accumulation turns these labelled counts into cumulative measurements; adjacent differences recover the increments wherever the walk includes them.

The six clean prefixes are 20, 40, 50, 250, 450 and 490. Together with the five female prefixes, they determine eight of the nine counts. The male-donkey count is absent from both registers, so it remains free. Adding the printed grand total of 550 supplies precisely that missing information and recovers the complete list. A smaller selection of nine measurements already suffices: all six clean prefixes, the third and fifth female prefixes, and the grand total.

This is a whole-list transfer of the cumulative-path rule. The reconstruction respects the source’s animal categories and the printed camel block; unnumbered young contribute no invented count. It also preserves the distinction between measurements and placements. A cumulative value such as 430 is an exact list result; assigning it a chronological comparison requires the separately declared dates and roles.""",
        "source_sections": [
            {"file": "File_58", "sections": ["§13.1", "§13.2", "§13.3", "§13.4"], "role": "Nine printed counts, category labels, three declared orders and total 550"}
        ],
        "C_steps": [1008, 1009, 1010, 1011],
        "key_claims": [
            "Clean plus female cumulative registers have rank 8; the grand total raises rank to9.",
            "The complete nine-count list is recovered without inventing quantities for unnumbered camel young.",
            "Source labels and orders precede accumulation; chronological placement remains a separate operation."
        ]
    },
    {
        "id": "basis_versus_compression",
        "title": "An exact change of coordinates, with its information cost",
        "text": """The Esau reconstruction identifies two complete ways to encode the same nine source counts. One uses selected clean and female cumulative measurements with the grand total. The other uses the complete clean and unclean walks. Their measurement matrices have determinants +1 and −1, respectively. Consequently, each encoding has an integer inverse: every count is recovered by exact additions and subtractions, without fractional corrections.

This result explains why a family can look different while retaining all its numerical information. Counts, successive landings and selected cumulative registers can serve as alternative coordinates for one labelled object. The calculation removes redundant reporting, but nine independent measurements still carry nine independent numeric inputs. It does not derive those source values from fewer free choices. Nor does an arbitrary integer measurement vector automatically describe admissible animal counts: nonnegativity and the source’s category meanings remain additional requirements.

Using every clean, female and unclean prefix gives fourteen measurements of rank nine. Five dependencies therefore follow from the measurement structure itself. Other equalities have a different status. The clean and unclean walks both reach 20 and 50, but through different animal rows; those agreements depend on the particular printed counts. Distinguishing structural dependencies from count-specific matches keeps the common grammar explanatory without multiplying the evidence supplied by equivalent descriptions.""",
        "source_sections": [
            {"file": "File_58", "sections": ["§13.1", "§13.2", "§13.3", "§13.4"], "role": "Source count vector and the incidence of each cumulative register"}
        ],
        "C_steps": [1011, 1012, 1013],
        "key_claims": [
            "Both selected nine-measurement bases have integer inverses; their determinants are +1 and −1.",
            "Coordinate recovery is distinct from reducing the number of independent source inputs.",
            "Fourteen registers have five structural dependencies; the additional 20/50 matches depend on the actual counts.",
            "Integer reconstruction does not by itself establish nonnegative admissibility."
        ]
    },
    {
        "id": "tishri_source_predicates",
        "title": "Which source conditions recover the Tishri ledger?",
        "text": """The complete Tishri ledger contains seven ritual categories and four species totals: 75 bulls, 18 rams, 176 lambs and 11 goats. Those margins alone leave the allocation among categories unresolved. A more informative reconstruction retains the source’s row templates and lets their seven amplitudes vary. Daily and Sabbath offerings contain lambs; the other templates retain their printed species ratios, including the complete Sukkot vector 70:14:98:7.

The four margins then provide four independent equations for seven amplitudes. Three additional source conditions close the system: Trumpets, Atonement and Eighth Day have equal amplitudes, giving two equations, and the daily contribution is 44 lambs. These recover the entire seven-row ledger. Removing daily 44 restores a Daily/Sabbath exchange; removing the festival equalities restores two allocation freedoms. The premises therefore have identifiable jobs rather than appearing as an undifferentiated collection of assumptions.

The seven-day Sukkot sequence gives a smaller version of the same result. Total 189, seven days and decrement 1 determine the totals 30 through 24. Retaining the uniform nonbull contribution 17 then gives the bulls 13 through 7. Other positive arithmetic bull sequences also sum to 70 if the decrement is free. Both reconstructions are conditional on numerical source templates and predicates; neither claims that a few totals independently generate the ritual system.""",
        "source_sections": [
            {"file": "File_58", "sections": ["§2.2", "§2.3", "§3.1", "§3.2", "§7.1"], "role": "Calendar and counting boundary, complete category ledger, and Sukkot daily sequence"}
        ],
        "C_steps": [1014, 1015, 1016],
        "key_claims": [
            "With seven source templates retained, the four margins have rank 4.",
            "Two equal-festival predicates and daily 44 raise rank to7 and uniquely recover the amplitudes.",
            "The retained templates contain numerical source information, including the full Sukkot vector.",
            "The daily sequence requires decrement 1 and the uniform nonbull 17 premise."
        ]
    },
    {
        "id": "nt_metric_coregistration",
        "title": "Separating the NT slot ratio, duration and placement",
        "text": """The NT lattice makes three kinds of information visible: a relation among slots, the duration of a slot, and the placement of the field. Let u denote one generation slot and w the Enoch-to-hinge radius. The supplied Key factors P=70/69 and E=25/23 send the Enoch radius to Jared one slot farther out and Adam six slots farther out. Their equations are (P−1)w=u and (E−1)w=6u.

These two equations contain only one independent condition. Since E−1 is six times P−1, both reduce to w=69u. They explain the relative named positions while leaving the unit’s duration free. The separately supplied BJ span of 2450 years across 35 slots supplies the metric: u=70 and w=4830.

Placement requires additional source information. In the existing display, BJ Creation at 3856 BC and Conquest at 1406 BC lie 55 and 20 slots from the birth hinge. Together those co-registrations recover the same 70-year unit and the 6 BC hinge. This is conditional identification inside an already constructed schematic field. It supplies neither an independent historical derivation of the dates nor two new independent proofs of the metric. Preserving the source’s birth, event and carrier roles prevents a successful coordinate recovery from silently becoming a claim about a different dated object.""",
        "source_sections": [
            {"file": "File_43", "sections": ["§§2.1–2.3", "§3.4"], "role": "Distinct names-span/interval states and the complete display-only ledger"},
            {"file": "File_54", "sections": ["§6", "§8.6", "§13.2"], "role": "Key fork; separate 2450/35-slot metric and the BJ co-registrations"}
        ],
        "C_steps": [1019, 1020, 1021, 1022],
        "key_claims": [
            "The one-slot P and six-slot E equations have rank 1 and imply w=69u.",
            "The source 2450/35 relation supplies the 70-year metric.",
            "The two existing co-registrations recover 6 BC and 70 conditionally within the schematic field.",
            "These are already fitted source relations, with birth, event and carrier roles retained."
        ]
    },
    {
        "id": "toledot_multiplicity",
        "title": "Keeping occurrence counts when sections are grouped",
        "text": """The Toledot table supplies eleven formula occurrences distributed across ten major sections. Its only repeated section is Esau: the formulas at Genesis 36:1 and 36:9 occupy positions 9 and 10 but belong to one major section. Terah is occurrence 6. There are therefore five formula occurrences before Terah and five after it, while the section count gives five earlier sections and four later ones.

The difference arises from the chosen measure. Counting sections assigns one unit to the Esau section; counting occurrences assigns it two. Retaining the ten section labels with multiplicities 1, 1, 1, 1, 1, 1, 1, 1, 2, 1 recovers the eleven-occurrence total and the five-on-each-side balance around Terah. The source need not abandon either grouping to explain why the counts differ.

This establishes a useful shared rule: an aggregation can preserve a source measure if its multiplicities travel with it. It does not follow that every operation on the original occurrence list survives the aggregation. In particular, the doubled Esau section still prevents index reflection from acting consistently on the ten section labels. The center established here is an occurrence-count center; it is not automatically a midpoint in elapsed years or a proof of semantic equivalence between paired headings. Those would require their own supplied measures and relations.""",
        "source_sections": [
            {"file": "File_70", "sections": ["Appendix B.1", "Appendix B.2"], "role": "Complete eleven-occurrence/ten-section table and distinct literary seam co-registrations"}
        ],
        "C_steps": [1023, 1024, 1025, 1026],
        "key_claims": [
            "Eleven formula occurrences map to ten sections by identifying Esau occurrences 9 and 10.",
            "Terah has occurrence arms 5/5 and section arms 5/4.",
            "Retained multiplicities recover occurrence measure without restoring reflection on the ten sections.",
            "Occurrence-count balance is distinct from a chronological midpoint or semantic equivalence."
        ]
    },
    {
        "id": "reflection_descent_comparison",
        "title": "When a reflection survives grouping",
        "text": """A reflection survives a grouping only when every pair of items grouped together has reflected images grouped together as well. This gives a direct test across complete source objects. On the eleven Toledot occurrences, index reflection sends position i to 12−i. The two Esau occurrences, 9 and 10, reflect to 3 and 2: Noah and Adam, which remain distinct sections. A single Esau section would therefore have two different reflected section images. Its multiplicity preserves the count but cannot resolve that ambiguity.

A formal repair would also group Adam and Noah, yielding nine classes. That calculation identifies the obstruction; the supplied source does not authorize the extra grouping. It is therefore excluded from the reconstructed literary object.

The primary NT path provides the positive comparison. Its 77 seventy-year intervals, from 5326 BC to AD 65, divide into eleven complete blocks of seven intervals. Reflection reverses the block order and preserves every selected boundary at indices 0, 7, …, 77. Each block represents 490 years, so the coarsening retains both the measure and the reflection. The separately displayed AD 65–135 extension is outside this primary path. These examples explain why retaining a total is weaker than retaining an action: source grouping, boundary selection and interval coarsening each require their own compatibility check.""",
        "source_sections": [
            {"file": "File_70", "sections": ["Appendix B.1"], "role": "Toledot occurrence order and the source-authorized Esau identification"},
            {"file": "File_43", "sections": ["§3.4"], "role": "Complete primary NT display intervals and separately printed final extension"},
            {"file": "File_54", "sections": ["§8.6", "§§10.1–10.3"], "role": "Primary NT/BJ partition and the 49/70/490 register"}
        ],
        "C_steps": [1025, 1026, 1027, 1028, 1029],
        "key_claims": [
            "A quotient carries reflection exactly when equal classes have equal reflected classes.",
            "Toledot's Esau class violates the criterion; an added Adam/Noah identification is diagnostic only.",
            "All eleven complete seven-edge NT blocks and all twelve selected boundaries are reflection-compatible.",
            "The 77-edge primary field excludes the separate AD 65–135 extension."
        ]
    }
]

for module in modules:
    module["word_count"] = len(re.findall(r"\S+", module["text"]))
    assert 150 <= module["word_count"] <= 250, (module["id"], module["word_count"])

packet = {
    "purpose": "Six reader-facing manuscript modules for integration into the Strategy explanation; draft only",
    "review_basis": "Independent C1008–C1031 review: 51 exact checks passed",
    "counting_rule": "Whitespace-delimited words in each text field; title and metadata excluded",
    "modules": modules
}

out = ROOT / "prep" / "manuscript_structural.json"
out.write_text(json.dumps(packet, ensure_ascii=False, indent=2) + "\n")
print(json.dumps({"path": str(out), "module_count": len(modules), "word_counts": {m["id"]: m["word_count"] for m in modules}}, indent=2))