BETA

Publishing a development

A development is a building or a neighbourhood with units inside. If you have already read publishing a listing, the flow is the same with one difference that changes the whole order.

The units are created BEFORE the development that groups them. This is not a style recommendation: the development references them, so if they do not exist yet there is nothing to reference. It is the number one mistake in this flow and it cannot be fixed going forwards — you have to go back.

The path

1 · each unit, separately         POST /properties             ← FIRST. They are normal properties
2 · the development grouping them POST /properties             ← references the ones from step 1
3 · check it                      POST /properties/verify      free
4 · publish                       PUT  …/publication/{portal}  ← and here it changes per portal
5 · status and takedown           same as a listing

Step 4 is not the same on the two portals, and it is the difference that changes your code the most.

ZonapropArgenprop
The unitstravel inside the development's publicationhave to be published already: the development takes them
How many calls you makeoneone per unit, and then the development's

On Argenprop each unit is a portal listing in its own right, and the development groups them by referencing them. A development cannot take a unit whose listing does not exist yet.

1 · The units first

Each unit is a normal property: it is created with the same POST /properties as any other.

⚠️ A unit is not a development. It has its own property type —apartment, house, plot— and it only inherits the development's if you get it wrong while building it. If that happens, the publish attempt flags it unit by unit, naming the type the portal does not recognise.

Store each one's code or identifier: it is what the development will reference them by.

On Argenprop, each unit also has to be published. Creating it here creates it on our side; the development can only take units that already have their listing on the portal. That is step 4.

2 · The development

Another POST /properties, with the development object and the list of the units from step 1.

Two things worth knowing beforehand:

  • A unit belongs to a single development. If it is already part of another, we tell you naming which one — we do not reassign it on our own initiative.
  • A development can be stored with no units and have them loaded later. What cannot be done is publishing it empty.

If you are a Mapaprop customer and generated your token from the app, this same step does something more: the development enters your Mapaprop account on its own once the photos finish processing, and its units are linked to it. It is stored, not published — publishing it is still step 4.

The development does not wait for units that have not arrived yet. It links the ones already there and returns the others in units.pending, with the reason for each one: this is normal when you upload the units and the development one right after the other, because each unit waits for its photos to finish uploading. To close them, retry the registration — only the missing ones are linked.

The detail of the mapaprop block is in Mapaprop.

3 · Check it (free, and here it pays more than ever)

The checker reviews the development and each unit, and it costs nothing. On Zonaprop, where publishing is a single call, a problem in one single unit can hold up everything; on Argenprop it does not hold up the development, but it leaves one unit out. In both cases finding out beforehand is free and finding out afterwards costs allowance.

The units have to exist BEFORE you check. The checker does not invent what is not there: if the development references a unit you have not created yet, it tells you so as not_found with the code it could not find — and concludes from that, rightly, that the development has no units.

So the ordering from step 1 applies to checking too, not just to publishing.

4 · Publishing

PUT …/publication/{portal} with the development. What happens to the units depends on the portal, and it is the only thing in this flow that is not the same on both.

On Zonaprop: one call, the parent and the units

The units travel inside it: you do not publish them one by one.

The allowanceit is one call, not one per unit. A development with 40 units costs the same as one with 2

A large development on Zonaprop is still one call

Nothing below applies to it: the units travel inside, so size does not change the number of calls.

On Argenprop: each unit first, the development afterwards

Each unit is a portal listing in its own right, so you publish it the way you would publish any listing —the same PUT …/publication/argenpropapi on the property— and only then do you publish the development, which takes them.

for each unit       PUT /properties/{paId}/publication/argenpropapi
and at the end      PUT /developments/{paId}/publication/argenpropapi
The allowanceone call per unit, plus the development's. A development with 40 units costs 41
The orderif you publish the development first, the units that do not have a listing yet come back as a failure, naming which one — the development does get published
Retrying is safetaking a unit that is already taken does not duplicate it nor cost you extra

Why we do not publish them for you internally. We evaluated it and it does not fit: publishing 30 listings inside a single call of yours takes longer than a request can last, and the failure mode would be the worst one —some published, others not, and you with no idea what got done—. Doing it one by one, each unit gets its own time budget and the flow scales without a ceiling.

And on both, the plan

The plandevelopments consume a plan of their own, different from a standalone listing's. Which one it is and how many your client has is what you see with the plans lookup

Depending on how many units it has, the portal may answer in one of two ways: confirming each listing, or confirming only the development and processing the units afterwards. What we return is explained in the development's response; the mode is not yours to choose.

How do I know everything I sent got published?

Do not count it by hand. There are two moments to find out, and the first one is the one worth using:

Before publishing, the checker already lists what we are going to send differently from how you declared it and what we are not going to send, with the reason for each. It is free, and it is the useful moment: you fix the object and check again without spending quota.

After publishing, the response brings what the portal accepted and keeps that same list. If a unit did not travel, it says which one and why.

A total that does not match what you sent is not, on its own, an error. It may be a unit we decided not to publish — the section that follows — or the mode where the portal processes afterwards and has not finished yet. Read the reason before retrying: retrying costs quota and is almost never the answer.

🔴 On Argenprop, a large development is published by REPEATING

Adopting the units one by one does not always fit in a single call —a call has a time limit— so the publish of a large development is naturally iterative:

you publish the development  →  it adopts what fits, and tells you how many are left
    you repeat                →  it picks up where it left off
        you repeat            →  …
            until none are left
Repeating is safeonly the missing ones are adopted. Nothing is duplicated and the development is not re-created
It is not an errorwhat did not fit comes back identified as not attempted, which is different from failed
What it costseach round is a few calls of your client's allowance. Do not loop blindly: read how many are left
When it is donewhen there are no units left to adopt

How many rounds: it depends on how fast the portal is answering that day, and it varies a lot. A development with few units usually closes in one; one with 50 may need on the order of 15 to 20.

If you are loading a large development for the first time, leave the loop running with a pause between rounds and watch the number of units left: it is the only indicator you need.

⚠️ This is only about the initial load. Afterwards, changing one unit's price or adding a single one is one call: we compare against what the portal already has and touch only the difference.

A unit that was sold: mark it, do not remove it

The development is the full inventory, not the offer of the moment. If you sold 3 out of 100 units, the 100 are still part of the development: you mark those 3 as sold and that is it. Do not remove them from the object.

Why it matters: the size of the development is a fact about the building, and the numbers we calculate — how many are left, how much has been commercialized — only make sense if they are all there. A development that went from 100 to 97 units lost the information that it sold 3.

And then each portal decides what to do with them. It is not the same everywhere: some portals have no way of showing a unit that is not available, so they would show it as if it were for sale. In those, we do not send it — and we tell you in the response, which one and why. The development's total stays the real one.

If the unit was already published, marking it does not take it down. Not sending it keeps its listing out of future updates, but the listing that is already up stays up: take down its publication like any other listing. They are two different things and it is worth doing both.

5 · Status, updating and takedown

Same as a listing: publishing again updates, the takedown is idempotent, and the status can be read for free on our side or by paying one call to the portal.

There is a window between publishing and being able to publish again, and here it is the longest one in the flow. With many units the portal does not finish on the spot: it acknowledges receipt and keeps working. While it does, it does not accept a second write of the same development — so an update sent right away is rejected.

It is not a failure and you did not lose the development: the first one carries on. We answer it as something temporary, with how many seconds it is worth waiting.

⚠️ The design decision this asks of you: if your flow publishes and corrects in the same burst, you have two paths and both are valid — wait and retry, or ask the portal whether it has finished before rewriting. Treating it as a definitive error is the only one that goes wrong. The contract of the publish response is in publish; the one for the check, in the status check.

Asking spends one call of your customer's quota; waiting is free.

With many units, asking will not help you here.

When the development is large the portal processes it asynchronously, and it finishes with the main listing long before it finishes with the batch. So if you ask, it will tell you that listing is ready while the units are still queued — and the next write keeps being rejected.

UnitsHow long until it accepted the next write
4~1 minute
20over 8 minutes (and the main listing reported ready at 3)
298 to 10 minutes

⇒ On a large development, wait and retry. The number we return in the response already takes the size into account: for a batch we suggest ten minutes, and for a standalone listing or a small development, one.

It is still a suggestion of ours, not a portal figure: there is no field that tells you the exact moment — the portal does not send it. If the same rejection comes back when it expires, nothing is broken: it is still processing, and retrying is safe.

The window also affects the TAKEDOWN, and there it surprises more. If you unpublish while the portal is processing, the takedown answers fine and the development may come back online when the portal finishes the work it had pending. It is not that the takedown failed: the previous publication was still in flight and completed afterwards. Same thing in taking it down.

⚠️ The development and its units have separate statuses. Ask about the development's through the development: asking about a unit as if it were a standalone listing will not tell you what you are after.

The three order mistakes, and how they look

If you…We tell you like this
send the development before the unitswe cannot find those units, with the code that does not exist
reference a unit that already belongs to another developmentwe name the development that has it
let the units inherit the development's typethe attempt flags each unit with the type the portal will not accept

All three come out before reaching the portal if you checked. If you did not check, they come out after spending the call.

Which portals support it

Publishing developments is not available on every portal, and on those where it is, your client needs to have the corresponding product contracted — which is different from the standalone listing one. The table is in Publication; if the portal does not support it, the attempt tells you without spending allowance.