With a town, it will produce 2 food and 7 commerce.
Not true.
End game financial towns produce 6 commerce unless riverside, in which case the corresponding financial farm produces 4 commerce, not 3. So let's start the comparison from the best possible case for cottages (they spring miraculously from the workers plow as fully developed towns with Taxation already in). This is very simple. What is more valuable: two food or three commerce? This question may be complicated by health caps, happy caps, available civics, individual city size and infrastructure, but at the stage in the game where Taxation is likely researched, there are plenty of interesting options for converting food into other resources and I would venture to guess that food in general will be more valuable.
However, that is possibly debatable. But the early game comparison for financial is simple. By the time a riverside cottage catches up in commerce to a riverside farm in 40 (normal speed) turns, they have both produced 120 commerce (with 40 bonus food from the farm). By the time a non-riverside cottage catches up in commerce to a non-riverside farm in 50 (normal speed) turns, they have both produced 150 commerce (with 50 bonus food from the farm). In both cases, the amount of commerce produced before the break even point is significant with the farm getting its commerce front loaded (better in general) relative to the cottage. More importantly, the food can be very valuable depending on each cities state, possibly being worth:
1-2 extra worked tiles in a city, more if a food saving building is in place
up to 60 hammers worth of a whip, more if a food saving building is in place
40 or 50 hammers on a pure one for one conversion of a worked non-farm grassland tile to an equivalent non-farm plains tile.
40 or 50 hammers directly into a worker or settler.
etc.
This varies from city to city obviously, but in general a new farm based city compare with a new cottage based city will:
1. Require fewer turns to reach its happy cap (and in the future fewer turns to reach a newly raised happy cap)
2. Require fewer worker turns to complete needed improvements (for example, early growth boosting farms that are converted to cottages later or chain irrigation passing farms for food bonuses in the cottage city will be needed in addition to the natural cottage and bonus improvements whereas in a farm city, such improvements are already part of the grand scheme)
3. Produce enough commerce to make it a viable contributor to the civilizations economy far quicker and maintain a commerce lead over the cottage counterpart for much much longer than the 40-50 turns indicated by the comparison between a single farm and a single cottage above.
4. Produce a higher output in terms of actual unit production (or building production if this is desired).
Also to address the cottage = warmonger improvement, this is not even close. Given sufficient commerce (and Aristocracy farms will easily provide sufficient commerce), warmongering is driven by production which is driven by food and hammers, not commerce. Simply put, the Aristocratic Farmer will be able to build a larger army than the cottager in the same amount of time.
Also, the 25% WW penalty is relevant. Regardless of Manors, that is a penalty to city states. The culture penalty is also a relevant hardship in that it slows the assimilation of newly acquired conquests and could imply smaller shared borders and thus potentially more distance to travel inside opposing civilizations borders, meaning potentially more supply costs and longer wars.
This is also ignoring the fact that cottages (still) are much more vulnerable to wars on their own soil. They are more reliant on workers, not less. They are more hurt by pillaging as developed cottages are harder to replace than farms, and given the decreased culture (city states), they potentially have smaller cultural radii and thus are more vulnerable to the potential of pillaging.
Also, just to point out one more thing about warmongering. For the Calabim it is very tough to justify swapping from Agrarianism to Conquest after Sanitation. With the Manors and sufficient health, they have an easy way to convert the 2 food from each farm to 1 hammer via unhappy citizens. This hammer is subject to multipliers. With conquest, their farms will only be producing one extra food which will be converted to 1 hammer via conquest. Assuming this works like workers and settlers (I can't say I've tried it so I'm not sure), this is not subject to multipliers, so results in fewer total hammers into the unit than Manors. And Manors are just the laziest conversion of food to hammers available to the Calabim and can easily be boosted by Sacrifice the Weak and/or Slavery. Plus these hammers can be spent on non units.