URL Segment Compression Part 3 — Global Token Budget

Continue the URL compression sequence. A path has major parts separated by / and minor parts separated by.. Base-compress a word by keeping its first and last character and replacing the middle with the number of omitted characters.
Each major may contain at most m output tokens, and the complete path may contain at most t output tokens. Every major must retain at least one token.
For this practice version, assign one token to every major first. Give the remaining global token budget to major parts from left to right, never exceeding m or that major's original minor-part count. Within a major with budget q, compress its first q-1 minor parts separately, concatenate all remaining original minor parts, and compress that concatenation as its final token.
The merge always uses original characters, not already-compressed text.
Practice sequence
Part 1: compress every minor part
Part 2: cap each major separately
Part 3: enforce a global token budget (current)

Function
compressGlobal(s: String, m: int, t: int) → String

Examples
Example 1
s = "alpha.beta.gamma/delta.echo.foxtrot/golf.hotel.india"
m = 2
t = 5
return = "a3a.b7a/d3a.e9t/g12a"
One token is reserved per major. The two remaining tokens go to the first and second majors, so their budgets are [2,2,1].
Example 2
s = "stripe.com/payments/checkout/customer.john.doe"
m = 2
t = 4
return = "s7m/p6s/c6t/c13e"
There are four majors and four global tokens, so every major receives one token and merges all of its original minor parts.

Constraints
1 <= m.
Let k be the number of major parts. k <= t <= the sum of all min(m, minorCount) values.
The input has no empty major or minor parts.
Every minor part has length at least 2.
The middle count may contain multiple digits.

The trick is two clean passes. First, parse into majors, each with its list of minor parts. Give every major one token, then walk left to right and hand out the leftover budget, capped by min(m, minorCount) minus the one already given. That gives you a budget q per major. Second, for each major, compress the first q-1 minors on their own, then concatenate all the remaining original minors and compress that string as one token. The pitfall is compressing first and merging after. Example 2 shows it: customer.john.doe becomes c13e, because the merged length is 15 and you omit 13 middle characters. Also watch multi-digit counts, and join tokens with. inside a major and / between majors. If the budget logic slips under pressure, StealthCoder can give you a working solution during the live OA.

Memorize the pattern. If you can't, run StealthCoder. The proctor sees the IDE. They don't see what's behind it.